Concepts In Bacterial Virulence (Contributions to Microbiology) [1 ed.] 3805577869, 9783805577861

Over the past few decades the increase in bacterial resistance has led to the search for novel antibacterial therapies a

227 21 16MB

English Pages 278 [287] Year 2004

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Concepts In Bacterial Virulence (Contributions to Microbiology) [1 ed.]
 3805577869, 9783805577861

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

ConUlbullonl. to

mャ」 ッ「セャッァケ

Editors A. SChmidt. H HOCmpo. セ (Krak6wI the.... Go.); Pike. R.N. (Clay",n)

T.; Molin. S. IセョケlH

181 Bacteriallnvasins:Holecular Systems Dedicated to the Invasion or Host Tissues Cambronne. E.O.; Schneewind, 0. (Chicago, Ill.) Signaling and Gene RegulatIOn

210 Bacterial Iron Transport Related to Virulence Bmun, V. (TUbingcn) 234 Pathogenicity Islands and Their Role in Bacterial Virulence and Survival Hochhul, 8.: Dobriru.h,

u,:

Hacker, J. (Wlirzburg)

2SS Horizontal and Vertical Gene Transfer: The Life History or Pathogens Lawrence, J.G. (Pittsburgh. Pa.) 272 Subject Index

Contenls

VI

Foreword

With the current volume of the Karger bouk series Contribulions /0 Microbiology, we attempt to summarize some of the most imponant virulence mechanisms in bacterial infectious diseases. In many cases the disease pathway begins with the invasion of the host and ends with the outbreak of physiological responses that may lead 10 severe complications and ultimately death. Over the years it has been shown that the interplay between pathogenic bacteria and the host is complex and finely balanced. The ability of successful pathogens to survive in an immunologically hostile environment is provided by a large armamentarium of virulence mechanisms, which includes bacterial factors thai evade, neutralize or counter the host defense systems, but also manipulate host homeostasis and nonnal cell functions. In order to give a comprehensive update, we were able to recmit some of the most eminent scientists in infectious diseases to give an overview of the most important recent findings in their fields. We hope that this volume provides a thought-provoking update on these important medical issues. Lund, May 2004

Wayne Russell Heiko Herwafd

VII

Toxins Russell W, Herwald H (eds): Concepts in Bacterial Virulence! Contrib Microbiol. Basel, Karger, 2005, vol 12, pp 1-27

Fundamentals of Endotoxin Structure and Function Russell E. Bishop Departments of Laboratory Medicine and Pathobio!ogy, and Biochemistry, University of Toronto, Toronto, Canada

In 1892, Richard Pfeiffer first defined endotoxin as a heat-stable toxic substance that was released upon disruption of microbial envelopes [I]. The toxicity is now known to be a consequence of the host inflammatory response, which appears to be optimally adapted for the clearance of most local infections. However, when severe infections become distributed systemically, the inflammatory response can lead to septic shock and death. Most of the early efforts to determine the signal transduction events that occur between the presentation of endotoxin to the myeloid cells of the immune system and the production of inflammatory cytokines have utilized lipopolysaccharide (LPS) from gramnegative bacteria [2]. The bioactive lipid A component of LPS is arguably the most potent of the substances that fit Pfeiffer's endotoxin definition, and lipid A has become synonymous with endotoxin. However, many other inflammatory mediators derived from bacteria can also be regarded as endotoxins, including peptidoglycan, the diacylglycerylcysteine moiety of bacterial lipoproteins, and bacterial nucleic acid signatures, to name only a few. The recent discovery that Toll-like receptor 4 (TLR4) is the lipid A inflammatory signal transducer has been followed by the identification of signal transducers for different inflammatory mediators [3, 4]. Coincident with these developments in endotoxin signaling has been the revelation that pathogenic gram-negative bacteria can modulate the structure of lipid A in order to evade detection by the host immune system. This article summarizes the recently elucidated pathways for the biosynthesis of lipid A in enteric bacteria, which provide a framework for understanding lipid A structure and function in all gram-negative bacteria. Readers are referred to the recent review of Raetz and Whitfield [5] for a more complete treatment of LPS structure and function that accounts for its diversity in more divergent organisms.

Overview of the Gram-Negative Cell Envelope

The cell envelope of gram-negative bacteria (fig. I) consists of the inner membrane (1M), the peptidoglycan (murein) and the outer membrane (OM) [5]. The 1M is a phospholipid bilayer, much like the plasma membrane of eukaryotic cells, and is permeable to lipophilic compounds. Numerous integral transmembrane a-helical proteins and peripheral membrane proteins are primarily responsible for transport, cell signaling and metabolic functions [6]. The 1M provides a topologically closed environment for the vectorial translocation of ions to generate a transmembrane electrochemical potential or proton-motive force that governs cellular energetics. Proteins synthesized with a cleavable amino-terminal signal peptide can be targeted for export across the 1M [7]. The periplasm is the gelatinous material between the OM and the 1M. It contains enzymes for nutrient breakdown as well as binding proteins to facilitate the transfer of nutrients across the 1M. Additionally, the murein sacculus in the periplasmic space is composed of alternating N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) sugars that are cross-linked by short peptide bridges [8]. The highly reticulated murein layer plays a crucial role in maintaining the cell's characteristic shape and in countering the effects of osmotic pressure. The murein is bridged to the OM by the abundant covalently bound murein lipoprotein, while numerous low-abundance non-covalently-bound lipoproteins are anchored to the inner leaflet of the OM and a few are anchored to the outer leaflet of the 1M. The OM is unique to gram-negative bacteria, and its role is to serve as a protective structure. The lipid arrangements of the OM are highly asymmetric. While phospholipids [70-80% phosphatidylethanolamine (PtdEtn), 20-30% phosphatidylglycerol (PtdGro) and cardiolipin] occupy the inner leaflet, LPS molecules pack against one another in a tight architecture in the outer leaflet of the OM [9]. Due to the low fluidity of lipid A hydrocarbon chains and the strong lateral interactions between LPS molecules, the OM bilayer is impermeable to lipophilic compounds and, thus, serves as an important permeability barrier for gram-negative bacteria [10]. To allow uptake of essential nutrients, the OM is studJed with trimeric (3-barrel proteins, knuwn as purins, which alluw diffusiun of solutes with a molecular weight below approximately 600 daltons. Additional (3-barrel proteins in the OM are adapted for the uptake of particular nutrients that cannot gain access through porins, and a few OM (3-barrel proteins function as enzymes [11]. One consequence of porins is that the OM is believed to lack any transmemhrane electrochemical potential. LPS is composed of three parts: the proximal, hydrophobic lipid A region, which anchors LPS to the outer leaflet of the OM, the distal, hydrophilic a-antigen repeats, which extend into the aqueous medium, and the interconnecting core oligosaccharide (fig. 2). The a-antigen and core sugars are not essential

Bishop

2

Transmembrane [3-barrel protein

c--

"

r"---

Outer membrane

v Lipoprotein



"""

"-

--

Mu rein --0--O-O-{>-O-0-C>-O--O-O-{>-O-0-Ct-O--O--0--O--:::>-O--0--O-O-{>-O-0-C>-O--O-0-D Peri plasmic space

f/Je,,,, peri plasmic proteins

Proton motive force

Inner membrane

Transmembrane «-helical protein

Fig. 1. Molecular organization of the gram-negative cell envelope. The OM is an asymmetric bilayer with an outer leaflet of LPS and an inner leaflet of glycerophospholipids Lipoproteins (GPL). The integral OM proteins are exclusively transmembrane セM「。イ・ャウN anchored to the OM inner leaflet can link the OM to the murein exoskeleton. The energytransducing 1M is a phospholipid bilayer that supports the proton motive force and contains transmembrane a-helical proteins. The periplasmic space is the region between the 1M and OM and contains numerous globular proteins.

for survival, but they provide bacterial resistance against various antirnkrobial agents including detergents and the membrane attack complex of serum complement [12]. Wild-type cells that produce O-antigen are termed 'smooth' due to their glossy colony morphology, while those that lack O-antigen are termed 'rough'. The term LPS formally applies only to the molecule that contains the

Endotoxin Structure and Function

3

a-antigen

Outer core

Inner core

EtN-P

Lipid A

2

Bishop

4

a-antigen polysaccharide, while molecules that lack a-antigen, as in the case of Neisseria, are more appropriately termed lipooligosaccharide or LOS. Lipid A is a target for the development of antibiotics and anti-inflammatory agents because it is both essential for survival and a potent inflammatory mediator.

TLR Signaling

When LPS is shed from the bacterial surface during infection, lipid A recognition in mammalian cells is mediated by the TLR4 signal transduction pathway [13, 14]. LPS is first recognized by the circulating acute phase LPSbinding protein (LBP), which then interacts with the glycosylphosphatidylinositolanchored CD14 on the surface of myeloid cells. Subsequent interaction with TLR4 and its associated factor MD2 initiates a cascade of signaling pathways that, in turn, elicit the production of cationic antimicrobial peptides (CAMPs), a variety of cytokine and chemokine molecules, and the costimulatory molecules that are expressed on the surface of antigen-presenting cells and further signal the presence of an infection to the cells of the adaptive immune system [15]. Upon activation, TLR4 recruits to its intracellular Toll-interleukin receptor homology region (TIR), the adapter protein MyD88, which associates by a homotypic protein-protein interaction with its own TIR domain (fig. 3). Another homotypic protein-protein interaction between the death domains of MyD88 and the interleukin-l receptor-associated kinase IRAK-I initiates the autophosphorylation of IRAK-I, which then associates with a signal transduction way station known as tumor necrosis factor-a (TNF-a) receptor-associated factor-6 (TRAF-6). An ubiquitin-conjugating enzyme complex is bound to TRAF-6 along with the TAK-l kinase complex, which is anchored by the TAB adapter proteins [3]. The pathway impinges on the master regulator of inflammation known as nuclear factor KB (NFKB), which activates transcription of inflammatory response genes. However, NFKB is normally sequestered in the cytoplasm in complex with its inhibitory subunit IKE. Proteolytic degradation of IKB enables NFKB lo migrale inlo lhe nucleus and aClivale inllammalory gene

Fig. 2. Structural organization of LPS. The most highly conserved region of the LPS molecule is the lipid A domain, which is an acylated and phosphorylated disaccharide of glucosamine. Assembly of lipid A is contingent upon the addition of the two 8-carbon Kdo sugars, which are the only essential components of the inner core. The inner core normally includes three 7-carbon Hep sugars and can be modified by the addition of phosphate and pEtN substituents. Outer core sugars provide the acceptor for a-antigen ligation, but tend to be composed of hexose sugars that differ between species. The a-antigens represent the most highly species-variable component of the LPS molecule.

Endotoxin Structure and Function

5

TNF-e

セ ⦅「 5) Dissolution

Fig. 1. The biofilm development cycle. Biofilm development is depicted as a general scheme involving attachment to the surface, fonnation of a tight association between bacterial cells and a surface, growth and intercellular adhesion allowing microcolony fonnation, maturation including EPS matrix development, and local dissolution leading to release of bacteria, which may eventually restart the cycle.

How do bacteria know that they are located in a biofilm? There is no doubt that cell density is an important factor that distinguishes the usually dilute suspensions of planktonic cells in water from the very cell-dense surface communities found where organic matter is abundant. One answer to the question therefore is: very high cell density. Another characteristic ofbiofilms and other types of surface-associated communities is the prevalence of internally heterogeneous environments and microenvironments, often generated and maintained by the presence of EPS. For the biofilm-associated bacteria this scenario is recognized as gradients of nutrients and stress factors. For planktonic cells such gradients rarely playa role. It is often argued that attachment to surfaces is the most important feature, and that surface-induced gene expression is therefore one of the key determinants ofbiofilm development. It should be remembered, however, that cellular contact with the substratum in a biofilm is a transient phenomenon (bul mosl likely imporlanl [or early gene aClivalion), which is quickly converted to a state where essentially all bacterial cells are located far above the surface in microcolonies or in EPS-embedded 'mushrooms'. In these entities it is difficult to imagine any bacterial sensing of the surface association as a physical signal. Thus, it seems that biofilm-associated bacteria must respond to the (1) very high cell density and (2) to the various positive and negative gradients. If it is assumed that bacterial evolution is mainly connected to the dominant life form of these organisms, and that bacteria in natural environments almost exclusively live an active proliferating life associated with surfaces (in biofilms), it is to be

Biofilms and Pathogenesis

115

expected that evolution has provided bacteria with properties that allow adaptation to life under high cell density conditions in environments with nutrient and antagonist gradients. This leaves the following issues as the major common themes for biofilm investigations related to the microbial capacity to develop mature, heterogeneously structured surface-associated communities: How are the specific structural features in a biofiLm created and maintained? Which functions are involved in the adaptation to high cell densities and nutrient gradients? How do biofilm bacteria evolve, and what are the major selective forces? In the following we will present an overview of the current understanding of microbial biofilm development and its clinical relevance in relation to two examples of gram-negative pathogens, Escherichia coli and Pseudomonas aeruginosa, for which the biofiLm lifestyle seems to be relevant dming the comse of infection.

E. coli

As the dominant facultative anaerobe of the normal human intestinal flora, E. coli remains harmlessly confined to the intestinal lumen. However, highly adapted clones have evolved the ability to cause a broad spectrum of diseases ranging from minary tract infection (UTI) and diarrhea to sepsis and meningitis [4]. Many of these infections are initiated by bacterial colonization of mucosal surfaces of the genitourinary, gastrointestinal or respiratory tracts. Successful establishment in the host depends on the ability to overcome host defenses and shear forces present at most of these surfaces. Since biofilm formation has also been suggested to be an ancient bacterial smvival strategy [5], it seems possible that at least a fraction of pathogenic E. coli clones have conserved or evolved the ability to enter a sessile lifestyle in multicellular biofilm communities in the host environment. Through investigations in recent years we now begin to realize that bacterial cell-cell interactions among E. coli cells on biotic and abiotic smfaces playa more significant role in pathogenicity than previously anticipated. It has therefore been of significant interest to clarify the mechanism(s) by which this organism colonizes surfaces and develops into substantial and robust biofilms.

In vitro Biofilm Development Since E. coli K-12 has been the workhorse bacterium for molecular biologists for nearly 50 years, standard laboratory strains became model organisms used in an approach to assign a developmental program to E. coli biofilms formed in vitro. A simple genetic screen was implemented utilizing 96-well

Reisner/Hoiby/Tolker-NielseniMolin

116

microtiter dishes as abiotic substrates for biofilm development in vitro, allowing large-scale isolation of mutants attenuated in biofilm formation under static conditions. Underlined by microscopic observations, the results ofthese initial studies were integrated in a developmental model for E. coli biofilm formation [6]. According to this model, E. coli K-12 utilizes flagella-mediated motility and type I pili to initiate early attachment processes. The major phase-variable outer membrane protein Ag43 was implicated in further development of microcolonies, and in agreement with the classical role ascribed to exopolysaccharides in stabilization of mature biofilms, the production of colanic acid was found to be required for the development of normal biofilm architecture in vitro. In subsequent similar approaches, additional factors have been found to affect biofilm formation of E. coli on abiotic surfaces in conventional growth media; however, only the effects of a few of them have been studied in detail [7]. The intracellular localization of most of the proposed effector proteins such as the disulfide bond formation catalyzing DsbA or the acetate kinase AckA suggests an indirect influence, possibly by altering expression, assembly or function of already implicated surface appendages and outer membrane proteins. The importance of others such as the stress-response sigma factor RpoS or the stringent response proteins RelA and SpoT might simply indicate the requirement for metabolic pathways and stress responses within the heterogeneous biofilms that are less important during exponential growth in suspension. Interestingly, the growth of E. coli K-12 biofilms in continuous hydrodynamic culture leads to the identification of biofilm-promoting factors, reflecting the reduced biofilm-forming capability ofK-12 lab strains under these conditions. An E. coli ompR234 mutant was isolated from the glass surface of a long-term continuous culture that was found to constitutively overexpress curli fimbriae [8]. The significantly improved biofilm formation phenotype was independent of flagella [9]. In 2001, Ghigo [10] discovered that conjugative plasm ids enhance biofilm formation on submerged Pyrex slides under continuous flow when the expression of conjugative pili is derepressed. Mutant analysis demonstrated that at least for plasmid F, functional conjugative pili are indeed necessary to obtain the observed induction. In a subsequent study, evidence was provided that the promotion of biofilm formation in the presence of the conjugative transfer genes of plasmid F is independent of flagella, type I pili or Ag43 synthesis [11]. As the biofilm lifestyle is thought to be fundamentally different from bacterial life in mixed suspension, major differences in gene expression were expected to be encountered upon switching from planktonic to biofilm growth. This view was confirmed by an experimental approach that used random chromosomal insertions of a promoterless lacZ reporter gene [12]. A large fraction

Biofilms and Pathogenesis

117

(38%) of 885 fusions was differentially expressed in a curli-promoted static E. coli K-12 biofilm when compared to planktonic cells. However, a recent microarray analysis of a biofilm formed by a wild-type K-12 strain under continuous flow indicated a more modest impact on global gene expression [13]. The transcript level of only 5.4 and 13.6% of the 4,290 protein-encoding genes was found to be significantly different as compared to expression in either exponential or stationary planktonic culture, respectively. It is unclear whether these drastically different results in terms of changes in global gene expression can be ascribed to the different strain background and/or the experimental setup. Due to the exclusive focus on K-12 strains in the vast majority of genetic studies, the relevance of the implicated factors for biofilm formation of nondomesticated E. coli isolates remains uncertain. For example, whereas the role of type I and curli fimbriae in the adherence of Shiga toxin-producing E. coli has been confirmed [14], a recent study suggests that the expression of colanic acid blocks adhesion of uropathogenic E. coli (UPEC) to inert abiotic surfaces [15]. Given the significantly elevated genome size of pathogenic E. coli as compared to K-12, determination of the diversity of molecular mechanisms used by the species E. coli in bacterial cell-cell interactions will necessitate the application of the already established molecular approaches at least to prototypic clinical E. coli isolates. Gastrointestinal Biofilms As a minority member of the normal flora of the large intestine in vertebrates, E. coli has to compete for nutrients with approximately 500 other indigenous species. In principle, successful coexistence can only be achieved by a growth rate that is at least equivalent to the washout rate from the intestine or by adherence to the intestinal epithelial cells [16]. Indeed, E. coli is capable of growing rapidly in intestinal mucus both in vivo and in vitro, whereas growth in luminal contents seems to be poor [17]. In addition, in situ hybridization experiments detected only separated single cells of commensal E. coli strains within the mucus layer but no bacterial cells associated with the epitheliuIIl [17, 18]. Thus, benign E. coli cells do not seem to be able to overcome the innate barriers that impede colonization in a healthy host and the natural lifestyle of these strains appears to be to reside and grow within the mucus layer almost exclusively as single cells. In contrast, each highly adapted E. coli clone causing diarrheal disease has evolved efficient ways to penetrate the mucus layer and stably adhere to the underlying epithelial cells even at intestinal sites normally not colonized by E. coli, such as the small bowel mucosa [19]. As for other mucosal pathogens, surface colonization by diarrheagenic E. coli is a prerequisite to initiate disease.

Reisner/Hoiby/Tolker-Nielsen/Molin

118

Not surprisingly therefore, the most useful phenotypic assay for the diagnosis and differentiation of diarrheagenic E. coli pathotypes is an adherence assay using monolayers of epithelial HEp-2 cells. Strikingly, the adherence pattern of members of two major pathotypes of diarrheagenic E. coli, enteropathogenic (EPEC) and enteroaggregative (EAEC) E. coli involves - in addition to binding to eukaryotic cells - apparent strong interactions between bacterial cells leading to three-dimensional structures typically observed in bacterial biofilms. EPEC develop a characteristic localized adherence pattern appearing as microcolonies on the surface, whereas EAEC appear to aggregate both on the surface as well as more distantly from the epithelium in a characteristic stacked-brick configuration [19]. Most importantly, similar biofilm-like adherence patterns have also been observed for both EPEC and EAEC in vivo. While the adherence to epithelial cells has been extensively studied, little information is currently available about the factors that trigger bacterial cell-cell adherence or the relevance of the size ofthese cell aggregates for pathogenicity [20]. Although the plasmid-encoded bundle-forming pili (BFP) of EPEC have been suggested to mediate interbacteria I interactions allowing formation of three-dimensional microcolonies on the surface ofepithelia [21], BFP-expressing EPECs were found to bind to epithelial cells rather than to already formed microcolonies. Interestingly, BFP are subject to morphological changes from thin to thick pili as infection proceeds, resulting in loosening and dispersal of the aggregates [20]. A bfpF mutant that was found unable to undergo this morphological change was significantly attenuated in virulence, indicating that formation and dispersal of microcolonies are both important for virulence. Likewise, plasmid-encoded thin aggregative adherence fimbriae were found to mediate the adherence and aggregation pattern of EAEC strains in vivo and in vitro [19]. Interestingly, the aggregative adherence pattern also requires expression of a secreted coat protein designated Aap (antiaggregation protein), which appears to promote dispersal ofEAEC on the intestinal mucosa by forming a protein capsule on the bacterial surface. Mutations in aap lead to increased aggregation and significantly reduced mucus penetration in vitro, indicating that bacterial cell-cell adherence has to be tightly controlled in order to be advantageous in the intestinal environment [22]. Nevertheless, a large fraction of EPECs and EAECs lack BFP and aggregative adherence fimbriae, respectively [19, 23]. Thus, E. coli clones seem to have evolved various divergent pathways to solve the same problem. Intracellular Biofilm-Like Pods in UTI The human urinary tract is usually a sterile system protected from the intestinal microflora by nonspecific resistance mechanisms that include phagocytosis, endotoxin-induced shedding of bladder epithelial cells, and the flushing

Biofilms and Pathogenesis

119

effect of urine flow. However, UTIs are considered to be the most common bacterial infections [24], with UPEC remaining the predominantly isolated species [25]. Generally, UPECs are thought to migrate from the gastrointestinal tract to the periurethral area where they eventually enter the bladder via the urethra [26]. Further transport into the kidneys may even enable an invasion into the bloodstream. Since intestinal E. coli clones are not equally able to survive within and colonize the urinary tract, UPECs are thought to be equipped with a variety of virulence factors including various adhesins offimbrial nature such as curli, type I pili, P, S, and FIC fimbriae [27]. These surface appendages bind to specific host cell receptor molecules and facilitate attachment of bacteria to specific epithelial cells they encounter during their transit [28]. However, despite the clear importance of cell-surface interactions during the course of infection, bacterial cell aggregates typical for biofilm formation have not been demonstrated on epithelial cells in vivo. Recent evidence suggests a novel role for biofilm-like cell-cell interactions during recurrent UTI. After artificial UTI infection of mice, Anderson et a1. [29] observed large pod-like bacterial cell aggregates within superficial cells of dissected bladders whereas un infected bladders appeared smooth. Bacteria within the pods had a uniform coccoid morphology, were interconnected by fibers and encased in a polysaccharide matrix. Although the presence of persistent E. coli in the bladder following acute UTI has been shown before, these large biofilm-like pods are observed after only 24 h of infection and represent a previously unrecognized intracellular microbial community and might playa role in the frequent recurrence of uncomplicated UTI (cystitis). However, the occurrence of these bacterial cell communities in human UTI has not yet been demonstrated.

Colonization ofIndwelling Devices For every artificial appliance placed in humans there is a corresponding microbial infection [30]. The crucial importance of biofilms associated with contamination of medical implant devices has been well established. Although E. coli has been found to adhere to implanted endotracheal tubes and contact lenses [6, 31], it is predominantly isolated from the surface of urinary catheters. Catheter-associated UTIs are indeed the most common among nosocomial infections. For example, 10-50% of patients experiencing short-term «7 days) urinary catheterization [32], and virtually all patients undergoing long-term (> 1 month) catheterization became infected [33]. During early stages of infection, E. coli is assumed to be present as a single species, whereas longer catheterization periods commonly lead to the formation of mixed communities of mainly gram-negative opportunistic pathogens,

Reisner/Hoiby/Tolker-NielseniMolin

120

including P aeruginosa, Proteus mirabilis, and Klebsiella pneumoniae [34]. Such E. coli-dominated biofilms formed on the luminal surfaces can reach more than 400 j.1m in height, are usually embedded in a polysaccharide matrix [35], and can contain minerals such as hydroxyapatite and struvite that crystallize at the biofilm-urine interface as a result of the elevated pH achieved by bacterial urease activity. Although symptoms are seldom associated with the infection initially, ultimate blockage of the inner lumen of the catheter and/or ascent of bacteria to the bladder and kidney manifest severe consequences for the patient if left untreated. Further support for a biofilm mode of growth after catheter colonization is derived from studies indicating that bacteria in these biofilms survive the urinary concentrations of antibiotics generated by standard treatment [36]. As a consequence, removal of the colonized device is the only efficient way to clear the infection. Given these complications generated by biofilms, several attempts have been made to prevent infection and bacterial colonization of catheters by incorporating conventional antibiotics or biocides such as silver oxide into the catheter material [34, 36]. Unfortunately, although the onset of bacteriuria could be delayed for several days with some catheter materials and treatments, most of these strategies were ineffective in preventing colonization [31]. A better insight into biofilm formation and ecology on catheters therefore appears to be required in order to identify more suitable and specific drug targets or to design more resistant catheters. It needs to be addressed whether initial colonization by E. coli supports a later establishment of other pathogens. Subsequent colonizers could attach to initial E. coli biofilms or benefit from provision of more suitable conditions in the local microenvironment such as changes of pH and nutrient supply. Interactions between different species during biofilm formation such as coaggregation might play an important role, as such phenomena have already been observed between lactobacilli and UPEC [37]. However, since standardized in vitro and in vivo models are crucial for obtaining any relevant information about virulence mechanisms, the lack of a nondestructive, longitudinal monitoring system is a major problem faced in indwelling-device-related biofilm research. A recently described mouse model of chronic biofilm infection that relies on biophotonic imaging ofbioluminescent reporter bacteria constitutes an appealing approach to overcome this bottleneck [38].

P. aeruginosa

P aeruginosa is an environmental microorganism found especially in freshwater and soil. In humans, P aeruginosa may cause a wide range of infections.

Biofilms and Pathogenesis

121

The most prevalent and severe chronic lung infection in cystic fibrosis (CF) patients is caused by mucoid, biofilm-forming P. aeruginosa, which has become endemic in CF patients [1]. CF is the most common congenital, inherited disease among Caucasian populations with an incidence rate of 1:2,500-1:4,500. The pathology of the lung infection, however, is similar in severe chronic obstructive pulmonary disease, where the number of patients is much higher.

In vitro Biofllm Development In contrast to the biofilm development for E. coli, which appears to be a case of relatively simple self-assembly processes in concert with surface association, P. aeruginosa is considered an example of a more elaborate biofilm developmental pathway involving several distinct steps of early and late maturation. Most of the work clarifYing this developmental cycle has been performed with reference strains - PAO I, P. aeruginosa 14 and PAK - and so far it appears that at least these strains share the major features of the biofilm developmental cycle. In particular, the highly structured P. aeruginosa biofilms (comprising 'mushrooms', 'towers', voids and water channels) observed under some conditions have been a challenge to molecular geneticists, and below we will briefly summarize the current understanding of how the development progresses and is controlled. It is first of all important to stress that structural biofilm development by P. aeruginosa appears to be conditional. The immediate environment is a key determinant of the eventual biofilm structure, illustrated by the finding that in flow chambers supplied with a citrate minimal medium P. aeruginosa fonns a flat biofilm, while in flow chambers supplied with glucose mjnimal medium it forms a heterogeneous biofilm with mushroom-shaped multicellular structures [39]. In a series of investigations, it was shown that the formation of the flat P. aenlginosa biofilm occurs via initial growth of sessile bacteria forming microcolonjes at the substratum, followed by expansive migration of the bacteria on the substratum, resulting in the fonnation of a flat biofilm [39]. Since biofilm formation by a P. aeruginosa pilA mutant (which is deficient in biogenesis of type IV pili) occurred without the expansive phase that results in discrete protruding microcolonies, it was suggested that the expansive mjgration of the bacteria on the substratum is type IV pili-driven, and that the shift may be induced by some sort of lirilltation arising in the initial microcolonies. The formation of the mushroom-shaped structures in the heterogeneous glucose-grown P. aeruginosa biofilm was shown to occur in a sequential process involving a nonmotile bacterial subpopulation, which formed the initial microcolonies by growth in certain foci of the bioItlm, and a migrating bacterial subpopulation, which initially fonned a monolayer on the substratum, and subsequently fonned the mushroom caps by climbing the microcolonies [40].

Reisner/Hoiby/Tolker-NielseniMolin

122

The nature of bacterial cell agglutinating factor(s) in very dynamic P aenlginosa biofilms is not lrnown at present. A role of alginate as a cell-to-cell interconnecting substance has been proposed previously [41], but recently it was concluded that alginate is not expressed at any significant level in such in vitro biofilms and therefore cannot be a key structural determinant under the defined conditions [42]. As we will see later, this situation is completely reversed in biofilms developing in some clinical cases, where alginate production appears to be essential for robust biofilm development. Some bacterial cell populations are apparently kept in the biofilm by substances that allow type IV pili-driven migration. Since twitching motility is powered by a mechanism involving extension, grip, and retraction of type IV pili [43], it is possible that type IV pili can playa role as cell-to-cell and cell-to-substratum interconnecting compounds. It has been reported that extracellular DNA may playa role as a cell-to-cell interconnecting substance in P aenlginosa biofilms [3, 44], and interestingly there is evidence that type IV pili bind to DNA [45]. Yet, other bacterial cell-to-substratum and cell-to-cell connections keep the pi/A mutant bacteria substratum-associated and agglutinated in the biofilms. Evidence is emerging that a novel type offimbriae may function as adhesin in P aeruginosa biofilms [46], and that certain exopolysaccharides may function as cell-to-cell interconnecting substances [Friedmann and Kolter, pers. commun.]. Such compounds could likely interconnect nonmigrating P aeruginosa populations. The apparent complexity of the biofilm developmental cycle of P aeruginosa has stimulated the search for genetic regulatory activities, and the findings of Davies et al. [47] that quorum-sensing control seems to be essential for normal biofilm formation was in accord with the characteristics of the process. In light of the current lrnowledge about the above-described steps ofbiofilm development for this organism it is, however, important to emphasize that so far no specific target for quorum-sensing control has been identified as relevant for these particular processes. It therefore remains to be seen whether quorum sensing is regulating any of the described process features such as bacterial cell-cell adherence, colony climbing or population differentiation. Chronic Lung Infections in CF CF patients are intermittently colonized with nonmucoid P aeruginosa strains for an average of 12 months before the infections become chronic, and the presence of mucoid strains and an antibody response is a sign of chronicity [48, 49]. The chronic P aeruginosa lung infections in CF patients is responsible for most of the morbidity and mortality of these patients [50], and this state of the infection constitutes a lung-associated biofilm [51, 52]. The biofilm is characterized by the mucoid phenotype of P aeruginosa producing an abundance of alginate [53]. In the conductive zone of the lungs the majority of the bacteria

Biofilms and Pathogenesis

123

stay inside the mucus and grow under anaerobic conditions using nitrate as electron acceptor [54]. Most of the bacteria are not located on the epithelial cells, but they induce an endobronchitis and endobronchiolitis without spreading to the blood or to other organs [54, 55]. In the respiratory zone of the airways, however, the environment is aerobic [56]. Foci of pneumonia in the alveolar tissue with extensive infiltration of polymorphonuclear leukocytes (PMNs) surround localized biofilms of P aeruginosa which are situated within the alveoles and alveolar ducts [55, 57]. The location and organization of the bacteria in these biofilms are similar to those observed in mucoid colonies and in sputum from CF patients with microcolonies of mucoid P aeruginosa [58]. High levels of antibodies are produced against alginate and other P aeruginosa antigens, but elimination ofthe infections is not accomplished [59], and the resulting persistent immune-complex-mediated inflammation is the major cause of the lung tissue damage [59]. The biofilm mode of growth is resistant to the patients' defense mechanisms and to antibiotic treatment [59] and is the major reason for the persistence of the infection lasting for more than 30 years in some patients.

Adaptation ofP. aeruginosa to CF Lungs The CF lung is a stressful environment for P aeruginosa, and, therefore, they have developed a range of survival strategies. When particles of >5 j.Lm containing bacteria are inhaled, they are deposited in connection with the gel phase of the mucus on the airway surfaces in the relatively small conducting zone of the central airways, which are covered by ciliated epithelial cells and coordinated movements of these cilia beating in the sol phase (=epithelial lining fluid) remove the gel phase of the mucus towards the trachea [56]. The gel phase of the mucus is produced by submucosal glands and goblet cells. In normal persons the effect of the cilia's beating (also named the mucociliary escalator) removes the mucus towards the trachea in this way rapidly (60 j.Lm/s) clearing the bacteria within 6 h [54, 60]. This clearance mechanism is the most important part ofthe noninflammatory defense mechanism of the respiratory tract. In CF patients, however, the basic defect of the CFTR protein leads to a reduced volume of the epithelial lining fiuid [60], and the mucociliary clearance of the bacteria is therefore greatly reduced, leading to robust bacterial growth [54] and recruitment ofthe inflammatory defense mechanisms (PMNs) [59]. When particles of2-5 j.Lm containing bacteria are inhaled, they are deposited in the much larger peripheral respiratory zone of the lungs without mucus or cilia, and the major defense mechanism are the alveolar macrophages, which belong to the inflammatory defense mechanisms [56]. In accordance, bronchoalveolar lavage studies on CF infants have shown that recruitment of the inflammatory defense mechanisms (dominated by the phagocytic cells, PMNs and macrophages) takes place when aspirated microorganisms are colonizing the lower respiratory

Reisner/Hoiby/Tolker-NielseniMolin

124

tract [61]. When PMNs and macrophages engulf bacteria there is a metabolic burst in the phagosomes leading to a release of reactive oxygen species, some of which are leaked to the environment [62]. These oxygen radicals induce killing, DNA damage and mutations in the bacteria [62, 63]. Oxygen radicals produced by the inflammatory response (PMNs) induce mutations in e.g. the mucA gene leading to the alginate production, which is characteristic for P aeruginosa biofilm infections in CF [64]. Alginate, on the other hand, is an oxygen radical scavenger [65] and provides mucoid P aeruginosa with protection against further DNA damage compared to nonmucoid strains [66]. Alginate can also make the bacteria resistant to phagocytosis by PMNs and macrophages [67]. Alginate production of P. aeruginosa biofilms in CF lungs, therefore, seems to be the major mechanism of adaptation permitting mucoid strains to persist in the hostile environment of oxygen radicals originating from the phagocytic cells of the inflammatory defense mechanisms. The lungs consist of the central conducting zone and the peripheral respiratory zone. When P aeruginosa grow in the peripheral respiratory zone (niche), the growth condition is comparable to growth in an aerobic or microaerophilic incubation chamber (5-20% oxygen). The respiratory zone is the area of the lungs where the venous blood becomes oxygenated in the dense capillary network of the alveoles, thus providing continuous culture conditions with nutrient and oxygen from the blood [56]. The central conductive zone of the respiratory tract (the bronchi), on the other hand, where P aeruginosa is located in sputum, is a completely different niche, since no oxygen is present in sputum [54]. Sputum consists mainly of dead PMNs and an abundance of released DNA [68] and leukocyte proteases [69] originating from PMNs in addition to mucus. In sputum the environment is anaerobic and the growth condition for P aeruginosa is comparable to a batch culture in the stationary phase. There is not so much blood supply of the conducting zone compared with the respiratory zone [56] and the bacteria are located inside sputum and not at the epithelial surface [54]. Under these conditions P aeruginosa may rely on anaerobic growth with N0 3 - as the electron acceptor [54]. In cases of infection with mucoid P aeruginosa cells, which dominates chronic infections, a pronounced antibody response against the bacteria is observed in connection with deteriorating lung function and poor prognosis. In contrast, the few CF patients colonized only with nonmucoid P aeruginosa have a low antibody response, and they maintain their lung function at the same nearly normal level similar to that ofCF patients without chronic infection [70]. The persistent PMN inflammation around P aeruginosa infection areas in the respiratory zone destroys the lung tissue of the infected foci of the lungs of the CF patients [71]. The alveolar macrophages in this zone [61], which migrate to the lymph nodes [56], are antigen-presenting cells, which are important for

Biofilms and Pathogenesis

125

initiating the antibody production of the B lymphocytes. Colonization of the conducting zone of the lungs, on the other hand, primarily leads to obstruction due to the abundance of mucus, and antibody production and lung tissue damage of the respiratory zone are normally not severe [54]. These observations suggest that severe respiratory failure in CF patients is caused by infection of the respiratory zone with mucoid P aeruginosa located in biofilms [55, 57]. Pieces of these biofilms are visible in gram-stained smears of sputum from CF patients [58]. Although the mucoid phenotype of P aeruginosa is characteristic for colonization of the respiratory zones in CF patients, nonmucoid variants of the same genotype are regularly present simultaneously in sputum [66]. The reason for this diversity has so far been obscure [58], but indications from in vitro investigations of stratified bacterial populations may be relevant for a better understanding of the phenotypical diversity of infectious P aeruginosa populations in CF lungs [57, 72-74]. In a population of lung-associated mucoid P aeruginosa, isogenic nonmucoid variants could represent a subpopulation of the original infecting cells (most likely not mucoid) occupying a niche in which mucoidy is not selectively favorable. Alternatively, the nonmucoid variants may be phenotypic revertants arising either as 'cheaters', benefiting from the alginate production of other bacteria within the biofilm, or as niche specialists in the anaerobic conductance zone. The fact that these variants seem to appear as individual bacteria outside the mucoid biofilm areas in sputum may indicate that they predominantly derive from the anaerobic zone.

Antibiotic Therapy Bacteria growing in biofilms are often much more resistant to antibiotics than planktonic cells of the same isolate. Minimal inhibitory concentration and minimal bactericidal concentration may be increased 100- to I,OOO-fold in old biofilms, whereas young biofilms are less resistant [75]. In contrast, planktonic bacteria released from such resistant biofilms are most often found to be as sensitive to antibiotics as the original planktonic cells [75]. Biofilm-induced resistance to antibiotics can be caused by several factors, such as slow growth, reduced oxygen concentrations at the base of the biofilm, penetration barriers e.g. binding of positive charges on the antibiotic molecules to the negatively charged alginate polymers, the presence ofl3-lactamase from the bacteria which cleaves and/or traps f3-lactam antibiotics and overexpression of efflux pumps [53, 76]. The increased resistance of biofilm bacteria usually results in the failure of antibacterial therapy with respect to eradication of the bacteria, but the antibiotic treatment regularly leads to temporary clinical improvement of the patient [53]. The development of traditional mechanisms of resistance to antibiotics occurs frequently in CF due to the intensive selective pressure provided by the

Reisner/Hoiby/Tolker-Nielsen/Molin

126

large amount of antibiotics used in these patients [53]. Mucoid and nonmucoid variants of the same strain are frequently simultaneously present in sputum but the nonmucoid variants are more resistant to antibiotics, possibly reflecting a higher antibiotic selection pressure outside the alginate biofilm [66]. The number of P aeruginosa in sputum may be as high as 108-10 1O CFU/ml. The high number of bacteria implies that mutations do occur in sputum. In addition, high frequencies (> 30%) of hypermutable P aeruginosa variants have been found in CF lung infection [77, 78], and the mutator strains (hypermutable strains) showing >20-fold higher mutation frequency than control strains [78] were also multiply resistant. The observations from P aeruginosa strains from CF patients showed the occurrence of a high frequency of hypermutable P aeruginosa, a high level of resistance to many antibiotics and, in the case of ciprofloxacin, several different mutations which increased over time [79]. In addition, mutations can be induced by means of oxygen radicals from PMNs, which in vitro leads to alginate production due to mutations in the mucA gene [64]. Furthennore, there is an antioxidant imbalance in the CF lung, which leads to oxygen radical damage [80]. Taken together, all these observations have led us to suggest that it is the chronic inflammation dominated by PMNs which induces a high level of mutations in P aeruginosa in the CF lungs and that the resistant mutants are then selected by the heavy use of antibiotics. These conventional resistance mechanisms are then added to the physiological resistance caused by the biofilm mode of growth in the CF lung.

Perspectives

There is an increasing documentation concerning the importance ofbiofilms in connection with microbial infections - in particular in relation to persistent infections of opportunistic pathogens. The detailed investigation of several microbial biofilms has produced interesting infonnation indicating that the multicellular life of bacteria may have its own genetic background that is controlled by baclerial inleraclions, which in some cases may resemble complex eukaryolic tissue development. One important question in relation to pathogenic bacteria is whether it is possible to extrapolate from these detailed in vitro observations and mechanisms to the conditions in the infected host. A word of caution is probably warranted: it is important to keep in mind that there is no indication of a consensus developmental program, and we therefore must resolve the individual biofilm pathways case by case. We also have strong indications that the in vitro biofilm conditions applied in the laboratory cannot be compared to those prevailing in the host, and it is therefore important to develop better model systems, if not perfonning the investigations in vivo. The genomic diversity of

Biofilms and Pathogenesis

127

bacteria is an additional complication; different isolates of the same species often behave quite differently from each other or when compared with reference strains or laboratory strains. We also have to keep in mind that simple molecular identification and characterization of various bacterial cell-cell interaction mechanisms only constitute the first step in an approach to interfere with cellcell interactions necessary for virulence. Since the overall physical strength and resistance of biofilms to shear force presumably playa critical role in vivo, a better understanding of the binding forces exhibited by the individual implied molecular factors is required to identify realistic drug targets. We now have some fundamental knowledge about the principles of bacterial life forms which seem to be important for a range of pathogens causing severe therapeutic problems in the clinic, and the technological and conceptual advances that have been made during the last 10 years ofbiofilm research should be applied with increased intensity in the investigations of infectious diseases. In particular, it will be important to establish the boundaries for our extrapolations from in vitro biofilm studies to the conditions prevailing in clinical cases, just as we must expand our investigation scenarios to encompass conditions which much better reflect what goes on in cases of suspected biofilm infections.

References

2 3 4 5 6 7 8

9

10 II 12

Costerton JW, Stewart PS, Greenberg EP: Bacterial biofilms: A common cause of persistent infections. Science 1999;284: 1318-1322. Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappin-Scott l-IM: Microbial biofilms. Annu Rev Microbiol 1995;49:711-745. Whitchurch CB, Tolker-Nielsen T, Ragas PC, Mattick JS: Extracellular DNA is required for bacterial biofilm formation. Science 2002;295:1487-1489. Donnenberg MS: Escherichia coli: Virulence Mechanisms of a Versatile Pathogen. London, Academic Press, 2002. Stood ley P, Sauer K, Davies DG, Costerton JW: Biofilms as complex differentiated communities. Annu Rev Microbiol 2002;56: 187-209. Davey ME, a'Toole GA: Microbial biofilms: From ecology to molecular genetics. Microbiol Mol Bioi Rev 2000;64:847-867. Ghigo JM: Are there biofilm-specific physiological pathways beyond a reasonable doubt? Res MicrobioI2003;154:1-8. Vidal a, Longin R, Prigent-Combaret C, Dorel C, l-Iooreman M, Lejeune P: Isolation of an Escherichia coli K-12 mutant strain able to form biofilms on inert surfaces: Involvement of a new ompR allele that increases curli expression. J Bacteriol 1998; 180:2442-2449. Prigent-Combaret C, Prensier G, Le Thi TT, Vidal a, Lejeune P, Dorel C: Developmental pathway for biofilm formation in curl i-producing Escherichia coli strains: Role of flagella, curli and colanic acid. Environ Microbiol 2000;2:450-464. Ghigo 1M: Natural conjugative plasmids induce bacterial biofilm development. Nature 2001; 412:442-445. Reisner A, Haagensen JAJ, Schembri MA, Zechner EL, Molin S: Development and maturation of Escherichia coli K-12 biofilms. Mol Microbial 2003;48:933-946. Prigent-Combaret C, Vidal a, Dorel C, Lejeune P: Abiotic surface sensing and biofilm-dependent regulation of gene expression in Escherichia coli. J Bacterial 1999; 181 :5993-6002.

Reisner/Hoiby/Tolker-Nielsen/Molin

128

13 14 15 16 17

18

19 20 21 22

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

Schembri MA, Kjaergaard K, Klemm P: Global gene expression in Escherichia coli biofilms. Mol Microbiol 2003;48:253-267. Cookson AL, Cooley WA, Woodward MJ: The role of type I and curli fimbriae of Shiga toxinproducing Escherichia coli in adherence to abiotic surfaces. IntJ Med Microbiol 2002;292: 195-205. Hanna A, Berg M, Stout V, Razatos A: Role of capsular colanic acid in adhesion of uropathogenic Escherichia coli. Appl Environ Microbiol 2003;69:4474--4481. Freter R, Brickner H, Fekete .I, Vickerman MM, Carey KE: Survival and implantation of Escherichia coli in the intestinal tract. Infect Immun 1983;39:686-703. Moller AK, Leatham Mp, Conway T, Nuijten PJ, de Haan LA, Krogfelt KA, Cohen PS: An Escherichia coli MG 1655 lipopolysaccharide deep-rough core mutant grows and survives in mouse cecal mucus but fails to colonize the mouse large intestine. Infect Immun 2003;71:2142-2152. Poulsen LK, Lan F, Kristensen CS, Hobolth P, Molin S, Krogfelt KA: SpatiaJ distribution of Escherichia coli in the mouse large intestine inferred from rRNA in situ hybridization. Infect Immun 1994;62:5191-5194. Nataro .IP: Diarrhoeagenic Escherichia coli; in Sussman M (ed): Molecular Medical Microbiology. London, Academic Press, 200 I, vol 2, pp 1463-1504. Nougayrede JP, Fernandes PJ, Donnenberg MS: Adhesion of enteropathogenic Escherichia coli to host cells. Cell MicrobioI2003;5:359-372. Giron .lA, Ho AS, Schoolnik GK: An inducible bundle-forming pilus of enteropathogenic Escherichia coli. Science 1991 ;254:71 0-713. Nishi .I, Sheikh .I, Mizuguchi K, Luisi B, Burland V, Boutin A, Rose DJ, Blattner FR, Nataro JP: The export of coat protein from enteroaggregative Escherichia coli by a specific ATP-binding cassette transporter system. .I Bioi Chern 2003;278:45680-45689. Trabulsi LR, Keller R, Tardelli-Gomes TA: Typical and atypical enteropathogenic Escherichia coli. Emerg Infect Dis 2002;8:508-513. Foxman B: Epidemiology of urinary tract infections: Incidence, morbidity, and economic costs. Am .I Med 2002;113:5S-13S. Ronald A: The etiology of urinary tract infection: Traditional and emerging pathogens. Am .I Med 2002; 113: 14S-19S. Sussman M: Urinary tract infections: A clinical overview; in Sussman M (ed): Molecular Medical Microbiology. London, Academic Press, 2001, vol 2, pp 1507-1514. Oelschlaeger TA, Dobrindt U, Hacker .I: Virulence factors ofuropathogens. CUfT Opin Uro12002; 12:33-38. Miihldorfer I, Ziebuhr W, Hacker .I: Escherichia coli in urinary tract infections; in Sussman M (ed): Molecular Medical Microbiology. London, Academic Press, 2001, vol 2, pp 1515-1540. Anderson GG, Palermo JJ, Schilling JD, Roth R, Heuser .I, Hultgren SJ: Intracellular bacterial biofilm-like pods in urinary tract infections. Science 2003;30 I: 105-107. Bisno AL, Waldvogel FA: Infections Associated with Indwelling Medical Devices. Washington, ASM Press, 1989. Donlan RM, Costerton JW: Biofilms: Survival mechanisms of clinically relevant microorganisms. Clin Microbiol Rev 2002;15:167-193. Mulhall AB, Chapman RG, Crow RA: Bacteriuria during indwelling urethral catheterization. .I Hosp Infect 1988;11 :253-262. Kunin CM: Care of the urinary catheter; in Detection, Prevention and Management of Urinary Tract Infections, ed 4. Philadelphia, Lea & Febiger, 1987, pp 245-298. Donlan RM: Biofilms and device-associated infections. Emerg Infect Dis 200 1;7:277-281. Ganderton L, Chawla .I, Winters C, Wimpenny .I, Stickler D: Scanning electron microscopy ofbacterial biofilms on indwelling bladder catheters. Eur .I C1in Microbiollnfect Dis 1992; II :789-796. Stickler DJ: Susceptibility of antibiotic-resistant gram-negative bacteria to biocides: A perspective from the study of catheter biofilms. .I Appl Microbiol 2002;92: 163S-170S. Rickard AH, Gilbert P, High NJ, Kolenbrander PE, Handley PS: Bacterial coaggregation: An integral process in the development of multi-species biofilms. Trends Microbiol 2003; II :94-100. Kadurugarnuwa JL, Sin LV, Albert E, Yu .I, Francis KP, DeBoer M, Rubin M, Bellinger-Kawahara C, Parr TR Jr, Contag PR: Direct continuous method for monitoring biofilm infection in a mouse model. Infect Immun 2003;71 :882-890.

Biofilms and Pathogenesis

129

39

40

41 42

43 44

45 46

47 48

49 50 51

52 53 54

55

56 57

58

59 60

Klausen M, Heydorn A, Ragas P, Lambertsen L, Aaes-J0rgensen A, Molin S, Tolker-Nielsen T: Biofilm formation by Pseudomonas aeruginosa wild type, flagella, and type IV pili mutants. Mol MicrobioI2003;48:1511-1524. Klausen M, Aaes-J0rgensen A, Molin S, Tolker-Nielsen T: Involvement of bacterial migration in the development of complex multicellular structures in Pseudomonas aeruginosa biofilms. Mol Microbiol 2003;50:61-68. Boyd A, Chakrabarty AM: Role of alginate lyase in cell detachment of Pseudomonas aeruginosa. Appl Environ Microbiol 1994;60:2355-2359. Wozniac OJ, Wyckoff no, Starkey M, Keyser R, Azadi P, O'Toole GA, Parsek MR: Alginate is not a significant component of the extracellular polysaccharide matrix of PA 14 and PAO 1 Pseudomonas aeruginosa biofilms. Proc Natl Acad Sci USA 2003; 100:7907-7912. Skerker lM, Berg HC: Direct observation of extension and retraction of type IV pili. Proc Natl Acad Sci USA 2001;98:6901-6904. Nemoto K, Hirota K, Murakami I(, Taniguti I(, Murata H, Viducic 0, Miyake Y: Effect ofVaridase (streptodornase) on biofilm formed by Pseudomonas aeruginosa. Chemotherapy 2003;49: 121-125. Schaik EJ, Irvin RT: DNA uptake in Pseudomonas aeruginosa: An additional type IV pilus function (poster). ASM Conference, Victoria, 2003. Vallet I, Olson JW, Lory S, Lazdunski A, Filloux A: The chaperone/usher pathways of Pseudomonas aeruginosa: Identification of fimbrial gene clusters (cup) and their involvement in biofilm formation. Proc Natl Acad Sci USA 200 I ;98:69 11-6916. Davies DG, Parsek MR, Pearson JP, Iglewski BH, Costerton JW, Greenberg EP: The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 1998;280:295-298. Hoiby N: Pseudomonas aeruginosa infection in cystic fibrosis. Relationship between mucoid strains of Pseudomonas aenlginosa and the humoral immune response. Acta Pathol Microbiol Scand Microbiol Immunol 1974;82:551-558. Johansen HK, Hoiby N: Seasonal onset of initial colonisation and chronic infection with Pseudomonas aenlginosa in patients with cystic fibrosis in Denmark. Thorax 1992;4712: 109-111. Hoiby N, Frederiksen: Microbiology of cystic fibrosis; in Hodson ME, Geddes 0 (eds): Cystic Fibrosis, ed 2. London, Arnold, 2000, pp 83-107. Singh PK, Schaefer AL, Parsek MR, Moninger TO, Welsh MJ, Greenberg EP: Quorum-sensing signals indicate that cystic fibrosis lungs are infected with bacterial biofihns. Nature 2000;407: 762-764. Middleton B, Rodgers HC, Camara M, Knox AJ, Williams P, Hardman A: Direct detection of N-acylhomoserine lactones in cystic fibrosis sputum. FEMS Microbiol Lett 2002;207/1:1-7. Hoiby N: Understanding bacterial biofilms in patients with cystic fibrosis: Current and innovative approaches to potential therapies. J Cystic Fibrosis 2002; I :249-254. Worlitzsch 0, Tarran R, Ulrich M, Schwab U, Cekici A, Meyer KC: Effects of reduced mucus oxygen concentration in airway Pseudonwnas aeruginosa infections of cystic fibrosis patients. J Clin Invest 2002;109:317-325. Baltimore RS, Christie CDC, Smith GJW: Immunohistopathologic localization of Pseudomonas aeruginosa in lungs from patients with cystic fibrosis - Implications for the pathogenesis of progressive lung deterioration. Am Rev Respir Dis 1989; 140: 1650-1661. Westh JB: Pulmonary Physiology and Pathophysiology. Philadelphia, Lippincott Williams & Wilkins, 200 I. Hoffmann N, Rasmussen TB, Jensen P0, Stub C, Hentzer M, Ciofu 0, Givskov M, Molin S, Johansen HI(, Hoiby N: A novel mouse model of chronic Pseudomonas aeruginosa lung infection mimicking cystic fibrosis. Infect Immun, submitted. Hoiby N: Pseudomonas aeruginosa infection in cystic fibrosis. Diagnostic and prognostic significance of Pseudomonas aeruginosa precipitins determined by means of crossed immunoelectrophoresis. A survey. Acta Pathol Microbial Scand 1977;262(suppl):3-96. Hoiby N, Johansen HK, Moser C, Song ZJ, Ciofu 0, Kharazmi A: Pseudonwnas aentginosa and the biofilm mode of growth. Microbes Infect 2001 ;3: 1-13. Boucher RC: An overview of the pathogenesis of cystic fibrosis lung disease. Adv Drug Deliv Rev 2002;54: 1359-1371.

Reisner/Hoiby/Tolker-Nielsen/Molin

130

61

62 63 64

65 66

67 68 69 70 71 72 73

74 75 76 77 78

79

80

Armstrong OS, Grimwood K, Carzino R, Carlin JB, Olinsky A, Phelan PO: Lower respiratory infection and inflammation in infants with newly diagnosed cystic fibrosis. Br Med J 1995;310: 1571-1572. Miller RA, Britigan BE: Role of oxidants in microbial pathophysiology. Clin Microbiol Rev 1997;10:1-18. Hull J, Vervaart p, Grimwood K, Phelan P: Pulmonary oxidative stress response in young children with cystic fibrosis. Thorax 1997;52:557-560. Mathee K, Ciofu 0, Sternberg C, Lindum PW, Campbell JIA, Jensen p, Johnsen AH, Givskov M, Ohman DE, Molin S, Hoiby N, Kharazmi A: Mucoid conversion of Pseudomonas aeruginosa by hydrogen peroxide: A mechanism for virulence activation in the cystic fibrosis lung. Microbiology 1999; 145: 1349-1357. Simpson J, Smith SE, Dean RT: Scavenging by alginate of free radicals released by macrophages. Free Radic Bioi Med 1989;6:347-353. Ciofu 0, Fussing V, Bagge N, Koch C, Hoiby N: Characterization of paired mucoid/non-mucoid Pseudomonas aeruginosa isolates from Danish cystic fibrosis patients: Antibiotic resistance, betalactamase activity and RiboPrinting. J Antimicrob Chemother 200 I;48:391-396. Simpson JA, Smith SE, Dean RT: Alginate inhibition of the uptake of Pseudomonas aernginosa by macrophages. J Gen Microbiol 1988;134:29-36. Shah PL, Scott SF, Knight RA, Marriott C, Ranasinha C, Hodson ME: In vivo effects of recombinant human DNase I on sputum in patients with cystic fibrosis. Thorax 1996;51/2: I 19-125. Goldstein W, Doring G: Lysosomal enzymes from polymorphonuclear leukocytes and proteinase inhibitors in patients with cystic fibrosis. Am Rev Respir Dis 1986;134:49-56. Pedersen SS, Hoiby N, Espersen F, Koch C: Role of alginate in infection with mucoid Pseudomonas aeruginosa in cystic fibrosis. Thorax 1992;47:6-13. Tiddens HAWM: Detecting early structural lung damage in cystic fibrosis. Pediatr Pulmonol 2002;34:228-231. Spiers AJ, Buckling A, Rainey PB: The causes of Pseudomonas diversity. Microbiology 2000; 146: 2345-2350. Wyckoff TJO, Thomas B, Hassett OJ, Wozniak OJ: Static growth of mucoid Pseudomonas aerugillOsa selects for non-mucoid variants that have acquired flagellum-dependent motility. Microbiology 2002; 148:3423-3430. Rainey PB, Travisano M: Adaptive radiation in a heterogenous environment. Nature 1998;394: 69-72. Anwar H, Strap JL, Costerton JW: Establishment of aging biofilms: Possible mechanism ofbacterial resistance to antimicrobial therapy. Antimicrob Agents Chemother 1992;36: 1347-1351. Stewart PS, Costerton JW: Antibiotic resistance of bacteria in biofilms. Lancet 200 1;358: 135-138. Oliver A, Canton R Campo P, Baquero F, Blazquez J: High frequency ofhypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 2000;288:1251-1253. Oliver A, Baquero F, Blazquez J: The mismatch repair system (mutS, mutL and uvrD genes) in Pseudomonas aerngillOsa: Molecular characterization of naturally occurring mutants. Mol Microbiol 2002;43: 1641-1650. Jalal S, Ciofu 0, Hoiby N, Gotoh N, Wretlind B: Molecular mechanisms offluoroquinolone resistance in Pseudomonas aeruginosa isolates from cystic fibrosis patients. Antimicrob Agents Chemother 2000;44:710-712. Wood LG, Fitzgerald DA, Gibson PG, Cooper OM, Collins CE, Garg ML: Oxidative stress in cystic fibrosis: Dietary and metabolic factors. J Am Coli Nutr 200 I;20/2: 157-165.

Soren Molin Molecular Microbial Ecology Group, BioCentrum-DTU Building 30 I, DTU DK-2800 Lyngby (Denmark) Tel. +4545252513, Fax +4545887328, E-Mail [email protected]

Biofilms and Pathogenesis

131

Enzymes Russell W, HelWald H (eds): Concepts in Bacterial Virulence. Contrib Microbial. Basel, Karger, 2005, vol 12, pp 132-180

Bacterial Peptidases Jan Potempaa, Robert N. Pike b Department of Microbiology, Faculty of Biotechnology, Jagicllonian University, Krakow, Poland, and Department of Biochemistry and Molecular Biology, University of Georgia, Athens, Ga., USA; b Department of Biochemistry and Molecular Biology, Victorian Center for Oral Health Sciences and CRC for Oral Health Sciences, Monash University, Clayton, Australia a

Enzymes that catalyze the hydrolysis of peptide bonds are referred to as proteases or peptidases. They are widely distributed in nature, where a variety of biological functions and processes depend on their activity. Regardless of the complexity of the organism, peptidases in general are essential at every stage in the life of every individual cell, since all protein molecules produced must be proteolytically processed and eventually degraded. Therefore, it is not surprising that throughout cellular life forms, genes encoding proteases occur at a relatively high frequency, ranging from 1.15% (Pirellula sp.) to 6.06% (Buchnera aphidicola) of the total gene count, with the average being about 3%. Among bacterial species which are pathogenic for humans, the number of peptidases known and putatively functional ranges from 9-15 in small genomes, such as those of the Mycoplasma spp. (1.45-2.07% of the total gene count) to 98 (2.64%) and 121 (2.85%) in genornes such as Pseudomonas aeruginosa and Escherichia coli, respectively. Fortunately, only a small fraction of the expressed peptidases in any pathogen impose a direct or indirect deleterious effect on their human host and may therefore be considered a virulence factor. With respect to the number of protease genes, the record in the microbial world goes to Bacillus cereus [179 potentially functional peptidase genes out of a total of 5,243 genes (3.99%)]. In comparison, only three times more functional protease genes have been identified in Homo sapiens (489 + 143 out of 23,531, 2.7% of the total gene count).

Classification of Peptidases Three major criteria are currently used to classify peptidases: (I) the reaction catalyzed, (2) the chemjcal nature of the catalytic site, and (3) the evolutionary relationship to other proteases, as revealed by the primary and/or tertiary structure of the protein. Based on the reaction they catalyze, peptidases are divided into two classes, comprising the exopeptidases and endopeptidases. The exopeptidases act only near the ends of polypeptide chains. Those acting at a free amino-terminus to liberate a single amino acid residue, a dipeptide or a tripeptide are referred to as aminopeptidases, dipeptidyl-peptidases, and tripeptidyl-peptidases, respectively. On the other hand, exopeptidases that cleave a single residue or dipeptide from a free carboxy-termmus are called carboxypeptidases and dipeptidyl-dipeptidases, respectively. Other exopeptidases are specific for dipeptides (dipeptidases), or the removal of termmal residues, either carboxy- or amino-terrrunal, that are substituted, cyclized, or linked by isopeptide bonds. Isopeptide bonds are peptide linkages other than those joining an a-carboxyl to an a-amino group. This last group is collectively referred to as the omega peptidases and is of particular importance for prokaryotic organisms producing nascent proteins that start with N-formylrnethionine at the beginning of their sequence, which needs to be removed. In contrast to the exopeptidases, endopeptidases preferentially hydrolyze peptide bonds in the inner regions of peptide chains, away from the termini. Typically, the presence of free a-amino or a-carboxyl groups has a negative effect on the activity of these enzymes, but it must be kept in mind that it is not unusual for an endopeptidase to have both exo- and endopeptidase activity. A subset of the endopeptidases, with activity limited to oligopeptides or fairly short polypeptide chains, are referred to oligopeptidases. According to the nature of their catalytic site, peptidases are divided into 6 types differing in their catalytic mechanism. The aspartic peptidases, sometimes incorrectly referred to as carboxypeptidases, have two aspartic acid residues invulved in the catalytic process. The cysteine-type peptidases (incorrectly called thiol peptidases) have a cysteine residue in their active center. The metallopeptidases use a metal ion (commonly zinc) in their catalytic mechanism. The activity of the serine-type peptidases depends on an active serine residue, while threonine-type peptidases utilize a catalytic threorune. The last group constitutes a numher ofpeptidases that cannot yet he assigned to any particular catalytic type. Among prokaryotic organisms, including pathogenjc bacteria, peptidases of all 6 catalytic types are common, although the frequency of their appearance is often strongly disproportionate (see following sections).

Bacterial Peptidases

133

A third way to classify peptidases is based on the evolutionary and structural relationship among enzymes, inferred from the comparison of amino acid sequences and/or tertiary structures. This method, introduced by Barrett et al. [2003], and currently implemented in the MEROPS database server (www. merops.ac.uk) [Rawlings et aI., 2004], is a powerful tool, allowing the logical classification of all peptidases, since the structural similarities within a family of peptidases commonly reflect important similarities in catalytic mechanism and other properties. However, in some cases, the classification is not fully consistent with three-dimensional structural data, as observed for the structurally distinct astacins and adamolysins, englobed in the same family M 12, or serralysins and matrixins, grouped into family MlO. This classification may even extend to assigning the biological function of an enzyme for which only the encoding DNA sequence is known. Therefore, the classification system briefly described below will be used here to discuss bacterial peptidases. The term 'family' is used to describe a group of peptidases in which each member shows an evolutionary relationship to at least one other, either throughout the whole sequence or at least in the part of the sequence responsible for catalytic activity. Each family is identified by an upper-case letter representing the catalytic type (A for aspartic type, C for cysteine type, M for metallo-type, S for serine type, T for threonine type, and U for unknown type), followed by a unique number. A family that contains deeply divergent groups is sometimes divided into subfamilies, identified by upper-case letters. Families are further clustered into clans. A clan contains all the present peptidases that have evolved from a single origin. It represents one or more families that show evidence of their evolutionary relationship, judged by similar tertiary structures, or when structures are not available, by the order of catalytic-site residues in the polypeptide chain and often by common sequence motifs around the catalytic residues. Each clan is identified by two letters, the first representing the catalytic type of the families included in the clan (with the letter 'P' being used for a clan containing families of more than one of the catalytic types: serine, threonine or cysteine). For the purpose of this review it is worth introducing a fourth classification of bacterial peptidases according to their role in pathogenicity. Pathogenicity, which is a term synonymous with virulence, is generally delineated as the ability of a bacterium to cause infection. Virulence factors represent either bacterial products or a strategy that contributes to virulence, which entails the pathogen to colonize the host, evade host defense mechanisms, facilitate dissemination, and cause host damage [Isenberg, 1988; Mekalanos, 1992]. In many respects, proteolytic enzymes produced by several pathogenic bacterial species fit into the category of virulence factors since they are directly involved in one or more of the processes listed above. Taking into account the numbers of peptidases produced by bacteria, relatively few can be considered sensu stricto as virulence factors. In

Potempa/Pike

134

this chapter we refer to peptidases, which preferentially target host proteins as 'primary virulence factors'. Many other peptidases are indirectly involved in pathogenicity, since they are indispensable for the expression of virulence factors per se. Such proteinases we call 'auxiliary virulence factors'. Finally, many other peptidases have well defined housekeeping functions. They do not harm the host either directly or indirectly, but are needed to withstand the stress of living in a hostile environment. We name them 'bystander virulence factors'.

Aspartic Peptidases

The MEROPS database currently (March 24, 2004) contains a total of 19,682 peptidase-related sequences and aspartic peptidases represent 6.3% of all peptidases, compared with 19.8% for cysteine, 30.2% for metallo-, 35.0% for serine, and 4.1 % for threonine peptidases. The aspartic peptidases are subdivided into six clans. Two clans (clans AC and AF) contain enzymes present only in the major domain of living organisms made up by bacteria. Bacterial peptidases also constitute a separate family within clan AD. They are represented by three archetypal enzymes: lipoprotein signal peptidase (LspA) often referred to as signal peptidase II (SPase II), a type IV prepilin peptidase and omptin. SPase II participates in prolipoprotein translocation through the cytoplasmic membrane of both gram-negative and gram-positive bacteria. With the exception of only three bacterial species, including Mycoplasma penetrans, Mycoplasma gallisepticum and onion yellows phytoplasma, the gene encoding a potentially functional protein has been found in all other species for which there is a completely sequenced genome (total 94). SPase II is a good example of a nonessential housekeeping enzyme, which, in the case of some pathogens, can contribute to their virulence. Apparently in Listeria monocytogenes, a gram-positive facultative intracellular human pathogen, temporally regulated expression of surface lipoproteins is critical for efficient phagosomal escape of L. monocytogenes. Mutants deficient in SPase II activity stayed entrapped inside the phagosomes of infected rnacrophages and have severely allenuated virulence [Reglier-Poupet et aI., 2003]. The gene encoding a potentially functional homologue of the type IV prepilin peptidase is strongly conserved amongst bacteria (clan AD, subfamily 24A), although not to the same degree as SPase II. The enzyme cleaves, among other substrates, the leader sequence from type 4 prepilins or prepilin-like proteins secreted by a wide range of bacterial species. Its activity is required for a variety of functions, including type 4 pilus formation, secretion of toxins and other enzymes through the type II protein secretion system in gramnegative bacteria, gene transfer and biofilm formation. In many regards,

Bacterial Peptidases

135

prepilin peptidase can be considered a housekeeping enzyme, but it contributes to the expression of well-defined virulence factors in several pathogenic species. In enteropathogenic E. coli, assembly of the type IV fimbriae known as the bundle-forming pilus (BFP) is dependent on the activity of the prepilin peptidase encoded by the bjjJP gene [Anantha et aL, 2000]. Biogenesis of BFP is required for autoaggregation and localized adherence to host cells and enteropathogenic E. coli mutants deficient in these surface appendages are nonvirulent in orally challenged human volunteers. Similarly, a knockout of the prepilin peptidase gene (pi/D) in Legionella pneumophi/a greatly impaired the ability of the bacterium to grow within amoebae and human macrophage-like U937 cells [Liles et al., 1999]. The mutant showed strongly attenuated virulence in animal models due to the malfunction of the prepilin peptidase-dependent type II secretion system operating inside the phagocytes [Rossier et al., 2004]. In the case of Vibrio cholerae, functioning of the extracellular protein secretion apparatus encoded by the eps gene is strongly dependent on prepilin peptidase activity. Deletion of the peptidase gene resulted in a dramatic decrease in cholera toxin secretion and abolished surface expression of the type 4 pilus responsible for mannose-sensitive hemagglutination [Marsh and Taylor, 1998]. In contrast to SPase II and the prepilin peptidase, which are good examples of auxiliary virulence factors, the plasminogen activating surface peptidase, Pia, of the plague bacterium Yersinia pestis is a paradigm for the primary virulence factor. The Pia surface peptidase resembles mammalian plasminogen activators in function and converts plasminogen to plasmin by limited proteolysis. At the same time, the Pia peptidase inactivates arantiplasmjn, a potent inhibitor of plasmin [Kukkonen et al., 2001], facilitating unrestrained activity of this broad-spectrum peptidase that in tum degrades fibrin and noncollagenous proteins of the extracellular matrix and activates latent procollagenases. This causes local damage of the connective tissue and enables the highly efficient spread of Y pestis from a subcutaneous site, where the pathogen is introduced by a vector bite, into the circulation [Sodeinde et al., 1992]. In addition, independent of proteolytic activity, the Pia peptidase mediates Y pestis adhesion to basement membrane and invasion into human endothelial cells, which may also contribute to dissemination of the bacterium in the host [Lahteenmakj et aI., 2001]. The Pia peptidase shares signjficant ammo acid sequence identity (about 50%) with the E. coli integral outer membrane peptidases, OmpT and OmpP, referred to as omptins. Since some serine protease inhibitors weakly affect OmpT activity and site-directed mutagenesis studies appeared to implicate Ser99 and His212 as the active site residues [Kramer et al., 2000], the omptins have been classified as novel serine proteases (family S18) [Rawlings and Barrett, 1994]. However, the crystal structure ofOmpT [Vandeputte-Rutten et aI., 200 I] followed

Potempa/Pike

136

by structure-guided site-directed mutagenesis [Kramer et al., 2001] proved that OmpT activity depends on the Asp83-Asp85 and Asp21 0-His212 residues. These residues are strictly conserved in all OmpT homologues described to date, including PgtE of the Salmonella sp., peptidase SpoA of Shigella flexneri, putative peptidases of Rhizobium loti, a new species of legume root nodule bacteria, plant pathogens of the Erwinia sp. and Agrobacterium tumefaciens, and of course OmpP and the PIa peptidase. It is assumed that these peptidases have a conserved fold, consisting of a 10-stranded antiparalJel l3-barrel that protrudes far from the lipid bilayer into the extracellular space with the catalytic site located in a groove at the extracellular top of the vase-shaped l3-barrel. Interestingly, activity of omptins is critically dependent on a specific interaction with lipid A of the LPS molecule [Kukkonen et al., 2004]. Omptins other than the PIa peptidase are typical housekeeping enzymes with their function/s not yet entirely understood. Nevertheless, they also seem to be implicated directly or indirectly in bacterial pathogenicity [Stathopoulos, 1998]. The presence of the omp T gene in clinical isolates of E. coli has been associated with complicated urinary tract disease [Webb and Lundigran, 1996], a notion supported by the observation that OmpT cleaves protamine, a highly basic antimicrobial peptide that is excreted by epithelial cells of the urinary tract [Stumpe et al., 1998]. Similarly, PtgE expression by Salmonella enterica may promote resistance to innate immunity by proteolytic inactivation of a-helical cationic antimicrobial peptides. On the other hand, SopA from S. flexneri, the causative agent of bacillary dysentery, cleaves the endogenous autotransporter IcsA, which has an essential role in the formation of actin tails in host cells, and therefore SopA might be indirectly involved in the actin-based motility inside infected cells [Egile et aI., 1997; Shere et al., 1997]. Among omptins only the PIa peptidase is a potent plasminogen activator. Interestingly, however, OmpT can be easily converted into the plasminogen activator by subtle mutations at surface-exposed loops. Such conversion may represent an interesting example of the evolution of a potent virulence factor from a housekeeping protein [Kukkonen et al., 200 I]. In the case of PgtE from S. enterica, the O-antigen ofLPS sterically prevents recognition of largemolecular-weight substrates, rendering plasminogen activator activity cryptic in this enteropathogen. The O-antigen repeats also prevent plasminogen activation by the Pia peptidase and, in this context, it is now clear why Y pestis lost the genetic locus involved in O-antigen synthesis [Kukkonen et al., 2004]. Collectively, it is apparent that the proteolytic activity of omptins contributes to virulence in a variety of ways. Their contribution ranges from bacterial defense and plasmin-mediated tissue infiltration to motility inside infected cells. Fortunately, they are produced by only a limited number of gram-negative bacteria which are pathogenic for plants and animals.

Bacterial Peptidases

137

Cysteine Peptidases The MEROPS database contains 3,897 cysteine-peptidase-related sequences (19.8% of the total sequences), which are divided into five phylogenetically related clans of proteins (CA, CD, CE, CF, and CH) and several families which are provisionally without a clan assignment. Bacterial peptidases are scattered among all of the clans except clan CH. It is a paradox, however, that although the bacterially derived cysteine peptidases, streptopain (SpeB) of Streptococcus pyogenes and clostripain from Clostridium perji-ingens were among the first proteolytic enzymes ever characterized, cysteine peptidases are underrepresented in prokaryotic organisms and show limited variation. Just one family (family C40) encompasses more than one third of the total cysteine peptidase count in prokaryotes (about 640 sequences). These enzymes are exemplified by dipeptidyl-peptidase VI from Bacillus sphaericus and murein endopeptidases (LytE and LytF) from Bacillus subtilis and represent typical housekeeping peptidases. Biochemically characterized enzymes have N-acetylmuramoyl-L-alanine amidase activity [Kuroda and Seikiguchi et aI., 1991; Moriyama et aI., 1996; Yamamoto et aI., 2003] and are involved in a peptidoglycan turnover. They are widespread among both gram-positive and gram-negative bacteria and genes encoding from I to 6 functional homologous are present in at least 70 bacterial species with completely sequenced genomes (out of94). No association with virulence has been reported for this group of peptidases.

Sortases (Family C60) Peptidases comprising the C60 family constitute a functionally and structurally related group of proteins expressed by all gram-positive species of bacteria. The prototypical enzyme, referred to as sortase A (SrtA), was first described in Staphylococcus aureus as an enzyme that is anchored in the plasma membrane and is responsible for covalent tethering of protein A to the cell wall [Mazmanian et aI., 1999]. It is now known that SrtA attaches a range of important surface proteins to the peptidoglycan component of S. aureus and many other gram-positive bacteria, including virulence-related microbial surface components recognizing adhesive matrix molecules (MSCRAMs). Substrates for SrtA are easily recognized by a carboxy-terminally located sorting signal made up by an LPXTG amino acid sequential motif, where X is any amino acid, followed by a hydrophobic domain composed of about 20 amino acid residues and a tail of positively charged residues. The hydrophobic domain and charged residues hinder polypeptide chain translocation through the plasma membrane, facilitating recognition of the LPXTG motif by SrtA. In a two-step transpeptidation reaction, sortase cleaves the LPXTG motif between the threonine and

Potempa/Pike

138

glycine residues and covalently attaches a polypeptide chain, via the carboxyterminal threonine, to the amino group of the pentaglycine crossbridge, thus tethering the protein to the cell wall. Although the structure of peptidoglycan crossbridging shows large variability in gram-positive bacteria, the mechanism of surface protein attachment is strictly conserved. A comparative genome analysis indicated that gram-positive bacteria frequently encode more than one sortase (up to 7 paralogues) and an even larger number of potential substrates (up to 40 per genome) with their characteristic LPXTG-type cell wall sorting motif or derivatives thereof [Comfort and Clubb, 2004]. In contrast, a single gene coding for a sortase and only one potential substrate have been identified thus far in only five gram-negative bacterial species. The sortases can be partitioned into 6 distinct subfamilies (5 in gram-positive and 1 in gram-negative bacteria) based on amino acid sequence. Members of each subfamily are suggested to recognize a discrete variation of the sorting motif [Comfort and Clubb, 2004]. In the bacterial species with more than one sortase, usually the SrtA-like molecule is responsible for tethering of most cell wall proteins in an organism, while additional sortase(s) have more specialized functions. For example, in the case of S. aureus, sortase B (SrtB) recognizes and anchors a protein known as IsdD, which is involved in heme iron transport [Mazmanian et aI., 2002, 2003]. This protein contains the NPQTN motif instead of the classical LPXTG sorting sequence exploited by SrtA, but otherwise the catalyzed reaction is identical. Also a protein, referred to as SvpA, which is anchored to peptidoglycan by SrtB of L. monocytogenes has the sorting motif, NAKNT, which is divergent from the one used by SrtA [Bierne et aI., 2004]. As in S. aureus, the genes encoding SrtB and its target, SvpA, are part of the same locus. In S. aureus, isd genes are regulated by iron and encode factors for hemoglobin binding and the passage of iron, in the form of a heme group, to the cytoplasm [Mazrnanian et aI., 2002]. Some ofthe six sortase genes encoded in the genome of Corynebacterium diphtheriae are required for biogenesis of the pilus. Assembly of the fimbriae involves the cleavage of pilin precursors at the classical sorting signal (LPLTG), or at an LAFTG motif, by two different sortases, which then further catalyze amide bond cross-linking of adjacent subunits or tethering to peptidoglycan [Ton-That and Schneewind, 2003]. This covalent attachment of adjacent pilin subunits has probably evolved in many gram-positive bacteria, since sortase genes in close association with pilin subunit genes with sorting signals were found in enterococci, streptococci, Actinomyces spp., and C. pelfringens. The NMR structure in solution of SrtA [Ilangovan et aI., 200 I] and the crystal structure ofSrtB [Zong et aI., 2004] from S. aureus are available, revealing an eight-stranded l3-barrel core structure with a helical subdomain at the aminoterminal end, which is unique among peptidases. The topology of the l3-barrel is

Bacterial Peptidases

139

identical in both enzymes with the critical cysteine residue (Cys 184 and Cys223 in SrtB and SrtA, respectively) located at the tip of the [37 strand. Initially, it was predicted that Cys184 and Hisl20 ofSrtA form a thiolate-imidazolium ion pair for catalysis [Ton-That et aI., 2002] as in the papain cysteine peptidases. However, pKa measurements for SrtB Cys184 and HisI20 residues refuted the involvement of the His residue in the transpeptidation reaction [Connolly et aI., 2003]. From the crystal structure of SrtB and conservation of the Arg233 (Arg197 in SrtA) residue it is apparent that a unique Cys-Arg catalytic dyad constitutes the foundation of the catalytic machinery of sortases. By exposing anchored proteins and polymeric structures such as fimbriae, the cell wall envelope of gram-positive bacteria can be considered to be a surface organelle maintaining contact between the microbe and its environment. It is now apparent that the assembly of these smface appendages is dependent on sortases. In this regard, sortases can be considered to be house-keeping enzymes. However, they are responsible for surface expression of acknowledged virulence factors, which mediate adherence to host tissues, host cell invasion, iron acquisition, and provide protection from assault by the formidable forces of the innate and acquired immune system. Therefore, sortases can be considered to be the classical example of an auxiliary virulence factor. Indeed, it was shown that sortase knockouts in various pathogenic bacteria, including S. aureus, Streptococcus mutans, L. monocytogenes, Streptococcus gordonii, and Streptococcus pneumoniae, have significantly attenuated virulence when tested in several different animal models. In this way sortase(s) are a very good target for the development of therapeutic inhibitors to fight gram-positive infections. Family C66: IdeS Peptidase (MAC Protein) A streptococcal protein (Mac) has been identified as a group A Streptococcus (GAS)-secreted protein of 35 kD with homology to the IT-subunit of Mac-I, a leukocyte [32 integrin. Mac binds to CD16 (FC)'RIIIB) on the surface of human polymorphonuclear leukocytes and inhibits opsonophagocytosis and production of reactive oxygen species, which resulted in significantly decreased pathogen killing [Lei et aI., 2001]. Later, the MAC protein was shown to be identical to the IdeS peptidase (IgG-degrading enzyme of S. pyogenes) [von Pawel-Rammingen et aI., 2002a, b], a previously unrecognized cysteine peptidase of S. pyogenes. The IdeS peptidase is an extremely specific enzyme, which exclusively cleaves the heavy chain ofIgG at the Gly237 residue in the hinge region. The enzyme is active in human plasma and its ability to interfere with Fc-mediated phagocytic killing has been demonstrated in a variety of bactericidal assays. These data collectively show that the IdeS protease contributes to evasion of the adaptive immune system by GAS by cleaving opsonizing IgG antibodies at the bacterial surface [von Pawel-Rammingen and Bjorck, 2003]. There is, however, a debate as to whether

Potempa/Pike

140

the proteolytic activity of IdeS (MAC protein) is absolutely necessary for interference with phagocytosis, which may only be dependent on molecular mimicry and the presence of the Arg-Gly-Asp amino acid motif in IdeS, which is involved in the interaction of the enzyme with the human integrins, av133 and au133 [Lei et aI., 2002; von Pawel-Rammingen and Bjorck, 2003]. The occurrence of orthologues of the IdeS peptidase is limited to a very small subset of the streptococci. In GAS, the enzyme occurs in two allelic variants among GAS serotypes, where the amino acid sequences of the variants differ from each other by about 15%. The only three homologues of the IdeS peptidase identified thus far are in the genome of Streptococcus equi (two genes) and in Streptococcus suis. One enzyme from S. equi was expressed and the recombinant protein was shown to possess the same activity as the IdeS peptidase [Lei et aI., 2003]. A distant homologue was also identified in the genome of Treponema denticola. The recombinant protein was expressed in E. coli and shown to have a nonspecific, general peptidase activity [Potempa, unpubl. data]. The activity ofIdeS depends on a thiolate-imjdazolium ion pair formed by Cys94 and His262, which act as the active-site residues as in the papain-like peptidases. These residues are conserved not only in the enzymes from S. equi and S. suis, but also in the T denticola homologue. However, the amino acid sequence is uruque and the crystal structure of the IdeS peptidase needs to be solved to delineate the relationshjp of the enzyme to other cysteine peptidases. Based on the present cumulative knowledge, it is apparent that the IdeS peptidase evolved to a primary virulence factor. It is also a good example of the possibility that bacteria may contain more peptidases than predicted from sequence alignments.

Clan CA All clan CA peptidases have a common fold motif, consisting of an ammotennjnal domain that is mostly a-helical and a carboxy-terminal domain featuring an antiparallel l3-sheet, with the Cys and His catalytic residues fonning a thiolate-inlidazolium dyad. However, it is also the most divergent and populous clan of the cysteine peptidases. The clan is divided into 12 families, of which bacterial peptidases are found only in 6. Two of these famjlies encompass exclusively bacterial enzymes that have apparently evolved as important virulence factors.

Family Cl: The Papain Family It is an evolutionary paradox that this major family of cysteine peptidases, exemplified by papain and mammalian cysteine cathepsins and encompassing more than 720 sequences, has only few representatives in bacteria. All together,

Bacterial Peptidases

141

only 47 homologues of papain have been identified, including 22 bacterial species with a completely sequenced bacterial genome. In this context, it is interesting to note that two Mycoplasma species, M. gallisepticum and M. penetrans, carry three and two copies of a gene encoding a potentially active papain homologue, respectively. However, among the genus Mycoplasma, these two species are the richest with regard to their peptidase gene count. Papain homologues occur predominantly in gram-positive species, the major representative being aminopeptidase C. This enzyme from Lactococci spp. has been thoroughly characterized [Vesanto et aI., 1994; Fenster et aI., 1997], and is also present in pathogens, but there are no reports that this peptidase or its homologues are involved in any aspect of bacterial pathogenicity. Family C2: The Calpain Family The protein fold of the peptidase unit for members of this family resembles that of papain. In mammals they are represented by calcium-regulated ubiquitous enzymes, but thus far only five highly diverged homologues have been identified in prokaryotes. The recombinant enzyme from Porphyromonas gingivalis, Tpr peptidase, was characterized as a general endopeptidase which also cleaves the bacterial collagenase peptide substrate. However, the enzyme has no collagenolytic activity [Bourgeau et aI., 1992] and there is no indication that the Tpr peptidase is associated with the virulence of this major periodontopathogen. Family CIO: The Streptopain (SpeB) Family The streptococcal cysteine peptidase was isolated and characterized in 1945 and was the second proteolytic enzyme after clostripain to be isolated from a prokaryote [Elliott, 1945]. For some time the identity of the peptidase was mistaken for the streptococcal pyrogenic toxin termed SpeB (streptococcus pyrogenic exotoxin B). The confusion ended when the entire genomes of several strains of GAS were sequenced, showing that SpeB and streptopain are the same protein. For historical reasons, however, streptopain is still very often referred to as SpeB. The enzyme occurs in two variants, which differ only in a single amino acid residue, glycine or serine, at position 164 from the aminoterminus of the mature enzyme. Most strains of S. pyogenes that are associated with severe invasive diseases express a Gly variant and therefore present an integrin-binding Arg-Gly-Asp motif at the surface-exposed loop. It was suggested that the ability of streptopain to bind integrins may be linked to the pathogenicity of these strains [Stockbauer et aI., 1999]. Despite a lack of significant sequence similarity, the crystal structure clearly indicates that streptopain belongs to the papain clan (superfamily) of cysteine peptidases. The mature peptidase portion has the two-domain fold

Potempa/Pike

142

characteristic of other papain-like enzymes, with an amino-terminal domain composed largely of a-helices and a carboxy-terminal domain based on a fourstranded antiparallel f)-sheet, with the catalytic dyad in the same topological orientation as in actinidin, a close relative to papain. In contrast to the peptidase domain, the profragment of streptopain has a unique fold. Willie an extended strand of the prosegment runs the full length of the active site cleft in a direction opposite to that of a natural substrate, thus blocking the major specificity pocket in the papain-like peptidase, in prostreptopain the inactivation mechanism relies on displacement of the catalytically essential histidine residue by a loop inserted into the active site [Kagawa et aI., 2000]. For more than 50 years, streptopain was recognized as a unique cysteine peptidase unrelated to papain or any other known peptidase. The first homologue of streptopain was identified in P gingivalis, a bacterium involved in the pathogenesis of human periodontal disease [Madden et aI., 1995], then another one from the same microorganism was purified and characterized [Nelson et aI., 1999]. This peptidase, referred to as periodontain, shows a strong preference for the degradation of unfolded polypeptide chains, with the human plasma proteinase inhibitor, ai-antitrypsin, being an important exception. This major inhibitor of human neutrophil elastase is very efficiently inactivated by cleavage in the reactive site loop [Nelson et aI., 1998]. Locally, this may lead to a loss of control of neutrophil peptidases and contribute to connective tissue damage. On the other hand, any direct role of periodontain in P gingivalis pathogenicity is obscure. The enzyme, together with its homologue, is probably involved in generating nutrients in the form of short peptides which are an indispensable source of carbon and energy for this asaccharolytic microorganism. The MEROPS database lists only three streptopain homologues, two in P gingivalis and one in the genome of Bacteroides thetaiotaomicron. However, closer analysis of partially finished bacterial genome sequences revealed that genes encoding potentially active streptopain-like peptidases are more widely spread. Three different homologues were found in the genome of Prevotella interrnedia, two in Prevotella ruminicola, and one in each of Tannerella forsythensis and Bacteroides fragilis. These genes encode either secreted or intracellular proteins. Significantly, the potentially secreted enzymes carry profragments with significant similarity to the proregion of streptopain. In the context of streptopain, which is very likely to be a virulence factor, it would be very interesting to elucidate the role of these streptopain homologues from other bacterial species. Streptopain is an outstanding example of a primary virulence factor with a very broad spectrum of activity. The list of pathogenetically relevant, biologically important proteins processed, activated, or otherwise altered by the enzyme is impressive. In vitro, streptopain cleaves the human interleukin-l f) (IL-l f))

Bacterial Peptidases

143

precursor to form bioactive IL-1 f3 [Kapur et aI., 1993a], processes the monocytic cell urokinase receptor [Wolf et aI., 1994] and degrades human fibronectin and vitronectin, two abundant extracellular matrix proteins engaged in maintaining host tissue integrity [Kapur et aI., 1993b]. In addition, streptopain activates latent human matrix metallopeptidases (MMPs), a process hypothesized to participate in the extensive soft tissue destruction observed in some patients with invasive streptococcal disease [Bums et aI., 1996]. Streptopain is able to cleave IgG molecules at the hinge region of the "i-chain, generating two Fab fragments and one Fc fragment [Collin and Olsen, 2000]. Interestingly, although streptopain can also cleave antigen-bound IgG, it does not affect antibodies bound to the bacterial surface through the Fc region [Eriksson and Norgren, 2003]. In this way, streptopain's ability to cleave off the Fc part of antigen-bound IgG contributes to the ability of GAS strains to escape opsonophagocytosis, while not interfering with the formation of a host-like coat oflgG immobilized on the bacterial surface through the Fc portion. This mechanism may significantly reinforce the defenses of S. pyogenes against attack by the adaptive immune response. In addition to streptopain, this deterrence system consists of (I) cell-wall-anchored surface proteins of the so-called M protein family, which binds IgG 'upside down' through the Fc fragment [Berge et aI., 1997]; (2) a secreted, highly specific endoglycosidase (EndoS) that targets conserved N-linked oligosaccharides on IgG [Collin and Olsen, 2000], and (3) the uniquely IgG-specific endopeptidase, IdeS (see family C66). Taken together, this system is very effective in protecting S. pyogenes against opsonin-dependent uptake and killing by professional phagocytes [Collin et aI., 2002]. Streptopain also seems to playa key role in shielding S. pyogenes from the innate immune system. The enzyme induces release of dermatan sulfate from the extracellular matrix resulting in the inactivation of antibacterial peptides [Schmidtchen et aI., 200 I] or directly eliminates the bactericidal potential of these peptides by degrading them [Schmidtchen et aI., 2002]. Finally, and possibly the most important role of streptopain in the pathogenicity of S. pyogenes is the ability of streptopain to directly release the potent peptide hormone, bradykinin, from high-molecular-weight kininogen. This release is not under the control of the host system [Herwald et aI., 1996]. Bradykinin released by bacterial pathogens has been shown to contribute to the dissemination of infection [Sakata et aI., 1996] and symptoms of sepsis and septic shock [Herwald et aI., 1998,2003; Tapper and Herwald, 2000]. Studies conducted with animal models confirmed the significant pathogenic potential of streptopain. The purified enzyme is lethal to mice [Gerlach et aI., 1983] and can cause myocardial necrosis when injected into rabbits, apparently due to its fibrinolytic activity [Kellner and Robertson, 1954]. Moreover, active immunization of mice with the purified streptopain elicits a protective response in a model of invasive

Potempa/Pike

144

disease, while mice injected with lethal doses of S. pyogenes were cured by a single injection of streptopain-specific inhibitor [Bjorck et aI., 1989]. Furthermore, experiments using a rat model of lung infection show that streptopain acts synergistically with either the streptococcal cell wall antigen or streptolysin 0 to augment lung injury [Shanley et aI., 1996]. This observation is especially intriguing in the context of the recent discovery that streptolysin 0 is the functional equivalent of the type III secretion system in gram-positive bacteria [Madden et aI., 2001] and invites speculation that in some circumstances streptopain may enter the host cell and act as an intracellular virulence factor. Taking into account the results of in vitro and ex vivo experiments, it is somewhat perplexing that the importance of streptopain as an indispensable virulence factor in vivo is still questioned. In one study, the importance of streptopain for the virulence of S. pyogenes has been demonstrated in a mouse model using isogenic strains with the streptopain gene inactivated by genetic manipulation [Lukomski et aI., 1997]. In the follow-up in vivo investigation, it was shown that streptopain helps S. pyogenes to resist phagocytosis [Lukomski et aI., 1998], contributes to soft tissue pathology, including necrosis, and is required for efficient systemic dissemination of the organism from the initial site of skin inoculation [Lukomski et al., 1999]. In stark contrast, in a welldesigned and executed study, Ashbaugh and Wessels [2001] proved that genetic inactivation of the streptopain gene did not significantly attenuate murine invasive infection, either after intraperitoneal or subcutaneous challenge. Also, in a model of necrotizing fasciitis, a streptopain mutant organism was found to be as effective in causing tissue damage, as the wild-type control strain [Ashbaugh et aI., 1998]. These results are in keeping with the clinical observation of an inverse correlation between disease severity and streptopain production in vitro by genetically related Ml Tl GAS isolates associated with invasive infection [Kansal et aI., 2000]. This paradox may be explained, at least partially, by the ability of streptopain to proteolytically remodel S. pyogenes surface proteins. Although this process is considered advantageous for bacteria [Rasmussen and Bjorck, 2002], two studies have suggested that the overexpression of streptopain results in nonspecific degradation of the antiphagocytic protein M and solubilizing of the C5a peptidase [Berge and Bjorck, 1995; Raeder et aI., 1998]. Together with degradation of secreted key virulence factors, such as superantigens (streptococcal pyrogenic exotoxins) [Kansal et aI., 2003], excessive production of streptopain may therefore decrease the pathogenicity of S. pyogenes. This hypothesis is further corroborated by the observation that streptopainnegative isolates have a survival advantage in vivo [Reader et aI., 2000] and the recent discovery that invasive M 1Tl GAS undergoes a stable phase shift to a phenotype expressing no streptopain, but instead a full repertoire of secreted

Bacterial Peptidases

145

proteins, which are apparently degraded by active streptopain [Aziz et aI., 2004]. This phenotypic phase shift may be related to the marked resurgence of severe, invasive and potentially fatal GAS infection, including the necrotizing fasciitis and streptococcal toxic syndrome observed during the last 20 years. The role of streptopain in GAS virulence confirms the ancient maxim that even for a bacterial pathogen too much of a 'good thing' can be bad. Indeed, S. pyogenes has developed its own system to regulate proteolytic activity and protect its surface-associated array of key virulence factors. Firstly, expression of streptopain is regulated at the transcriptional level [Heath et aI., 1999]; secondly, streptopain is produced as an inactive zymogen, which undergoes an autocatalytic, multistep activation process assisted by the bacterial surface [Liu and Elliott, 1965a, b; Collin and Olsen, 2000; Chen et aI., 2003], and thirdly, in vivo, the pathogen can coat its surface with the broad spectrum peptidase inhibitor, cx2-macroglobulin (cx2M) immobilized through interaction with the peptidoglycan-anchored protein, G-related cx2M-binding protein (GRAB). Bound to GRAB, cx2M protects protein M, and possibly other surface proteins, from being cleaved by streptopain [Rasmussen et aI., 1999]. In this context, it is very interesting to note that S. pyogenes retains some of the streptopain enzyme displays associated with the bacterial cell surface, where the enzyme displays laminin-binding activity [Hytonen et aI., 2001]. Taking into account the mechanism of peptidase inhibition by cx2M, it is tempting to speculate that the immobilized form of streptopain preserves proteolytic activity even in the presence of a high concentration of this inhibitor. Such a feature may be particularly useful in soft tissue infections where the experimental and epidemiological evidence strongly implies that streptopain plays a critical role in promoting infection [Svensson et aI., 2000].

Family C47: The Staphopain Family At present, this family is limited to the Staphylococcus genus. Staphopain occurs in two variants, apparently reflecting the duplication of an ancestral gene. S. aureus expresses both variants, referred to as staphopain A and staphopain B, which share about 47% identity at the amino acid sequence level of the mature enzymes. The single staphopain of Staphylococcus epidermidis is related to staphopain A (75% identity) [Dubin et aI., 2001; Oleksy et aI., 2004]. On the other hand, a gene encoding a close relative of staphopain B has been cloned from Staphylococcus warneri, while a cysteine peptidase similar to the staphopains was purified from the growth medium of Staphylococcus simulans biovar staphylolyticus [Donham et aI., 1988; Neumann et aI., 1993]. Both staphopains are processed from large precursors, but so far only the crystal structure of the mature staphopains is available [Hofmann et aI., 1993; Filipek et aI., 2003]. Remarkably, despite the low sequence similarity to

Potempa/Pike

146

papain-like peptidases, the tertiary structure of the staphopains resembles the overall fold of papain. The reciprocal relationship present between the staphopains apparent at the amino acid sequence level is also mirrored at the genetic level. The staphopain A gene (scpA) occurs in a bicistronic operon (scpA), in which it is followed by a gene (scpB) encoding a staphopain A-specific inhibitor. On the other hand, the staphopain B gene (sspB) is part of the tricistronic operon sspABC, where sspA and sspC encode the V8 protease and an inhibitor specific for staphopain B, respectively [Rzychon et aI., 2003a, b]. The staphopain inhibitors, ScpB and SspC, termed staphostatins, have similar folds and apparently the same mechanism of target peptidase inhibition although they share less than 20% sequence identity [Rzychon et aI., 2003a, b, Dubin et aI., 2003]. Nevertheless, they are uniquely specific; ScpB affects only staphopain A activity, while SspC exclusively inhibits staphopain B, without any cross-reactivity. In some cases, the reactivity of the inhibitor does not extend to the orthologous enzyme from other staphylococcal species [Dubin et aI., 2004]. Apparently, evolution has hand-tailored these inhibitors to control the activity of the coexpressed enzyme. Interestingly, staphopains are secreted, while staphostatins are intracellular proteins, suggesting that they function as so-called threshold inhibitors protecting cytoplasmic proteins from any prematurely folded peptidases [Rzychon et a!., 2003a, b]. The genetic assembly of peptidase and inhibitor genes in cotranscribed, cotranslated tmits provides the means for very efficient elimination of active staphopain from the cytoplasm. The extracellular activity of S. aureus is also the subject of multilevel control. All secreted peptidases, including both staphopains are coordinately regulated at the transcriptional level by an accessory gene regulator operon (agr) in a cell density-dependent manner [Janzon et aI., 1989]. This regulation is fine tuned by direct, strong repression of the transcription of the stpAB and sspABC operons by SarA, the product of the staphylococcal accessory regulator (Sal) locus [Chan and Foster, 1998; Lindsay and Foster, 1999; Ziebandt et aI., 2001]. Additionally, this regulatory system is indirectly affected by the alternative sigma factor (J'B [Ziebandt et aI., 2001] and probably by several SarA-like transcriptional factors. Collectively, this highly complex network of gene regulation assures the precisely coordinated synthesis of extracellular proteins, including staphopains and other peptidases. In the case of the proteinases, the regulation of their activity does not stop at the transcriptional level. Aureolysin, the V8 peptidase (glutamylendopeptidase I) and the staphopains are secreted as proenzyme forms and activated in a cascadelike marmer. It is well documented that aureolysin activates the zymogen of the V8 peptidase, which in turn cleaves pro-staphopain B [Drapeau, 1978; Rice at aI., 200 I]. Indeed, pro-staphopain B can be expressed in the zymogen form in E. coli

Bacterial Peptidases

147

and activated in vitro by the V8 peptidase (1. Potempa, unpubl. data). In contrast, the means by which pro-staphopain A processing/activation occurs is obscure and nothing is known as to whether this pro-enzyme is inactive or which proteinase is responsible for its processing. Tight regulation of staphopain expression, together with that of other acknowledged virulence factors, including toxins and adhesins, may be considered as indirect evidence of their importance for the survival of S. aureus in vivo. This association has revitalized interest in staphylococcal extracellular peptidases as markers of pathogenicity, a subject which has been neglected for many years. Unfortunately, the results of recent investigations using animal models of staphylococcal infection are contradictory and confusing. Firstly, it was shown that a mutant strain deficient in the V8 peptidase was severely attenuated in virulence in mouse abscess, bacteremia and wound infection models [Coulter et aI., 1998]. However, the reduced virulence of this mutant was apparently due to a polar effect on the expression of the sspB gene encoding staphopain B, located downstream of the V8 peptidase gene (sspA) in the same operon [Rice et aI., 200 I]. Indeed, this assumption was confirmed using a S. aureus strain with the staphopain B gene eliminated by means of genetic manipulation [Shaw et aI., 2004]. In this study it was shown that only the sspB gene knockout strain, but not the metalloproteinase (aureolysin) and staphopain A-deficient mutants were attenuated in the skin abscess model. However, these results were not confirmed in a model of septic arthritis in mice. The inactivation of any of the peptidase genes did not affect the frequency or severity of joint disease, indicating that, at least in this model, staphopain B does not act as virulence factor [Calander et aI., 2004]. Taken together, the role of staphopains in the physiology and virulence of staphylococci is obscure, but stringent conservation of the stpA and sspB genes among S. aureus strains, as well as preservation of the stpA-like gene among coagulase-negative staphylococcal species, implies that their function is important for staphylococcus survival in vivo. Amongst the bacterial proteinases, staphopains are unique with regard to their secretion as zymogens and activation by limited proteolysis. In this respect they resemble streptopain from S. pyogenes. In addition, for an as yet not understood reason they are tightly regulated both at the transcriptional and posttranslational levels. At the protein level their activity is released in a cascade pathway unique among bacterial species and then is further controlled by highly specific inhibitors.

Family C39: Bacteriocin-Processing Peptidase Bacteriocins are antimicrobial peptides produced by microorganisms belonging to different bacterial taxonomic branches and used by microorganisms for biological warfare and communications [Eijsink et aI., 2002]. One type of

Potempa/Pike

148

these peptides is posttranslationally modified (class I lantibiotics), while a second type does not contain modified amino acids (class II nonlantibiotic bacteriocins). Both classes are ribosomally synthesized in the precursor form. In most nonlantibiotic peptides and some lantibiotic peptides, the amino-terminal extensions are composed of a very characteristic leader sequence termed the doubleglycine-type leader, which is cleaved after the second glycine, concomitant with export carried out by members of a specific family of dedicated ATP-binding cassette (ABC) transporters. The amino-terminal domain of these transporters, absent in other ABC transporters, contains conserved cysteine and histidine residues operating as the catalytic dyad. Also, other residues, including the glutamate and aspartate residues which participate in peptide bond hydrolysis by papain-like peptidases, are strictly conserved in this portion of the molecule, which apparently has a canonical fold characteristic of papain [Havarstein et aI., 1995]. The peptidase domain, together with a central hydrophobic integral membrane domain and a carboxy-terminal cytoplasmic ATP-binding domain, constitutes the dedicated transport machinery which recognizes substrates and removes leader peptides while trans locating them across the cytoplasmic membrane. In addition to bacteriocins, the ABC transporters are used to translocate peptide pheromones [Michiels et aI., 2001]. Bacteriocin-processing peptidases are widespread amongst both grampositive and gram-negative bacteria and constitute the second most numerous family of cysteine peptidases in prokaryotes (after family C40). None has been implicated as a virulence factor. On the contrary, as peptidases which are indispensable for the maturation of bacteriocins, they can be utilized in expanding applications using bacteriocins as natural food preservatives [Riley and Wertz, 2002]. Family C51: D-Alanyl-Glycyl Endopeptidase Representatives of this family have thus far only been found in the three bacterial species, S. aureus, S. epidermidis, and S. pyogenes. The enzymes are phagederived and can degrade the cell wall envelope. Autolysins LytN and LytA from S. aureus possess a D-alanyl-glycyl endopeptidase as well as N-acetylmuramylL-alanyl amidase activity, which is contained within the amino-terminal portion of the polypeptide chain [Navarre et aI., 1999]. None of these autolysins has been implicated in virulence. Conversely, it has been suggested that they may be used to counter antibiotic-resistant staphylococcal infections [Fischetti, 2003]. Family C58: The YopT Peptidase Family Bacterial pathogens share common strategies to infect and colonize animal and plant host [Staskawicz et aI., 2001]. One system, widespread among gramnegative pathogens, referred to as the type III secretion system [Cheng and

Bacterial Peptidases

149

Schneewind, 2000; Comelis and Van Gijsegem, 2000] directly delivers different classes of proteins to the host. These proteins, now collectively termed type III effectors, mimic, suppress, interfere, or modulate host defense signaling pathways. Their sole function is to enhance pathogen survival, proliferation and dissemination and therefore may be considered to be primary virulence factors. The structural scaffold to dispense type III effectors is conserved but 'delivered goods' are custom designed to serve the particular needs of a given pathogen. This is exemplified by the YopT peptidase [Comelis, 2002] and its homologues from Yersinia spp. and plant pathogens, including Pseudomonas syringae [Axtell et aI., 2003], which, despite sharing the same fold and catalytic mechanism, target a different set of substrates inside host cells. In addition to the YopT peptidase onthologues, an overlapping set of pathogens has adopted a cysteine peptidase with a different fold and evolutionary origin (clan CE) [Orth, 2002] as the type III effectors. The YopT peptidase is one of six proteins called Yop effectors (YopH, YopE, YopJNopP, YopONpkA, YopM, and YopT) injected into the host cell by the Yersinia type III secretion system [Juris et aI., 2002]. They function in concert to thwart the host immune system. YopT itself exerts a cytotoxic effect in mammalian cells when delivered by the type III secretion system [Iriarte and Comelis, 1998]. This effect is due to proteolytic cleavage ofposttranslationally modified Rho GTPases by the YopT peptidase [Shao et aI., 2002]. Apparently the YopT peptidase specifically recognizes prenylated Rho GTPases and executes a proteolytic cleavage near their carboxy-termini [Shao et aI., 2003b]. This leads to the loss of the carboxy-terminal lipid modification on these GTPases, resulting in their release from the membrane and irreversible inactivation. Globally, this causes a disruption of the actin cytoskeleton, exerting a powerful antiphagocytic effect and thus protecting the pathogen from being killed by phagocytes. AvrPphB is an avirulence (Avr) protein from the plant pathogen P syringae that can trigger a disease resistance response in a number of host plants. The crystal structure revealed that the topology of the catalytic triad (Cys-His-Asp), together with other structural features, resembles that for papain-like peptidases, particularly staphopain [Zhu et aI., 2004]. AvrPphB has a very stringent substrate specificity and apparently exerts only a single proteolytic cleavage in the Arabidopsis serine/threonine kinase PBSI [Shao et aI., 2003a]. It is suggested that the cleavage product is recognized by RPS5, a member of the class of R proteins that have a predicted nucleotide-binding site and leucine-rich repeats. In a resistant host these molecular events induce a hypersensitive response. The avr genes of the YopT family are common amongst plant pathogens as well as symbiotic plant bacteria and multiple Avr proteins are found in a single

Potempa/Pike

150

Pseudomonas strain. They all function as specific peptidases targeting different substrates in the plant host or possibly cleaving the same substrates at different positions, generating signals detected by distinct R proteins. It is speculated that the large number of YopT-like proteins found in plant pathogens may reflect coevolutionary pressures in which the evolution of a new R protein in the host that detects the cleavage products of a given peptidase selects for a pathogen with new protease variants [Axtell and Staskawicz, 2003; Zhu et aI., 2004]. Clan CD This clan was recognized based on a conserved sequential motive His-Glyspacer-Ala-Cys encompassing the catalytic His-Cys dyad present in caspases, peptidases involved in apoptosis and cytokine activation (family 14), gingipains (family 25), plant and animal legumains, processing proteinases (family 13), bacterial clostripain (family II), and separase, a proteinase required for sister chromatid separation during anaphase (family 50) [Chen et aI., 1998]. The additional common feature of all these enzymes is a substrate specificity dominated by a specific PI residue recognition, which is asparagine (Iegumain), lysine (Kgp), arginine (Rgp, clostripain, and separase), or aspartic acid (caspases). Although crystal structures are only available for caspases and one gingipain, it is expected that representatives of other families in the clan will also have a similar fold. The hallmark of this fold is a six-stranded parallel l3-sheet in the middle of the molecule sandwiched by three ex-helices on each side [Eichinger et aI., 1999]. Out of the five CD clan families known so far, three are found in bacteria. Family Cll: The Clostripain Family Clostripain was identified and partially purified in 1937 from the culture filtrate of Clostridium histolyticum. The enzyme was then characterized as a cysteine peptidase that is strictly specific for Arg-Xaa (Xaa stands for any amino acid) peptidyl bonds. The mature, active clostripain is a noncovalent heterodimer derived from an inactive precursor through the autocatalytic removal or a 9-residue linker peptide [Wille et aI., 1996, 1994]. At least 16 clostripaiIll onthologues homologues were identified in microbial genomes, most of エィ・セ in Clostridium spp. [Labrou and Rigden, 2004]. None of them was ever implicated as a virulence factor in clostridial infections. On the contrary, clostripainl is a very useful enzyme in technology, both in sequence analysis and in enzyJ 1 matic peptide synthesis [Gunther et aI., 2000]. Family C13: The Legumain Family Mammalian asparaginyl endopeptidase (AEP) or legumain is a recently identified lysosomal cysteine peptidase belonging to clan CD. To date it has been

Bacterial Peptidases

151

shown to be involved in antigen presentation within main-histocompatibilitycomplex (MHC) class II-positive cells and in proprotein processing [ShirahamaNoda et aI., 2003; Manoury et aI., 1998; Sarandeses et aI., 2003]. Genes encoding potentially active legumain homologues have thus far only been found in a few bacterial species, including Caulobacter crescentus, P aeruginosa, Pseudomonas putida, P syringae, Xanthomonas axonopodis, and Xanthomonas campestris. Their function awaits elucidation.

Family C14: The Caspase Family Caspases are important players in the programmed cell death of multicellular organisms ranging from humans to sponges [Wiens et aI., 2003]. Comparative genomic studies have provided evidence which indicates that the eukaryotic apoptotic system emerged by acquisition of several central apoptotic effectors, including caspases, from a-protobacteria as a consequence of mitochondrial endosymbiosis [Koonin and Aravind, 2002]. Therefore, it is not surprising that homologues of caspases, referred to as paracaspases and metacaspases [Aravind and Koonin, 2002], are abundant in diverse bacteria, particularly those with complex development, such as Streptomyces, Anabaena, Mesorhizobium, Myxococcus, and a-protobacteria. The role of these ancient enzymes in bacterial physiology is obscure. Family C25: The Gingipain Family So far gingipains have only been found in P gingivalis, the major pathogen of adult onset periodontal disease. They are represented by the products of three genetic loci conserved amongst clinical and laboratory strains of P gingivalis, one (kgp) encoding a lysine-Xaa peptide bond-specific endopeptidase (gingipain K, Kgp) and two others, rgpA and rgpB, which are arginine-Xaa-specific enzymes (Arg-gingipains, Rgps) [Curtis et aI., 1999; Potempa et aI., 1995]. The nascent translation products of gingipain genes undergo complex proteolytic processing and posttranslational modifications [Veith et aI., 2002]. In the case of Kgp and RgpA, initial polypeptide chain fragmentation is necessary for assembly of a noncovalent complex composed of the catalytic, hemoglobin-binding and hemagglutination/adhesin domains [Potempa et aI., 2003]. This complex is either anchored to the outer membrane through a glucan moiety attached to the carboxy-terminus of the domain derived from the carboxy-terminal portion of the nascent product, or released into the growth media in the nonglycated form. RgpB lacks the additional hemoglobin-binding and adhesin domains, but still undergoes complex modification consisting of the autoproteolytic removal of the profragrnent and either truncation at the carboxy-terminus (the secreted form of the enzyme) [Mikolajczyk et aI., 2003] or glycosylation at the carboxyterminus, the latter allowing RgpB to form an association with the cell envelope

Potempa/Pike

152

[Veith et aI., 2002]. Collectively, gingipain activity constitutes at least 85% ofthe general proteolytic activity produced by P. gingivalis [Potempa et aI., 1997]. In every respect, gingipains can be considered to be primary virulence factors for P gingiva lis-dependent initiation and/or progression of periodontal disease. As peptidases, they target a large set of disease-relevant substrates which can be directly associated with the clinical hallmarks of the disease [Potempa et aI., 2000]. Due to the large number of substrates it targets, gingipain activity is also thought to provide this asaccharolytic organism with nutrients. However, gingipains are certainly broad spectrum peptidases. Actually, in many cases they act with the precision and sophistication of the tailored host peptidases, mimicking their function. The best example of how P. gingivalis can manipulate the host is the use of the gingipains to affect the major proteolytic cascades of coagulation, complement activation, fibrinolysis and kinin generation [Imamura et aI., 2003]. The coagulation cascade is targeted at several levels by Rgps, which convert factor X, factor IX, protein C and prothrombin to active peptidases by limited proteolysis, thus mimjcking the action of host enzymes [Imamura et aI., 1997, 200 Ia, b; Hosotaki et aI., 1999]. In the case of factor X activation, this functional mimicry additionally involves enhancement of the Rgp-converting activity in the presence of phospholipids and Ca2 +, two critical cofactors of the normal coagulation cascade [Imamura et aI., 1997]. The factor X activation is very efficient, with the catalytic potency in some cases matchjng that of natural activators. In this context it is worth emphasizing that gingipains are not controlled by host inhibitors, in stark contrast to the clotting factors. In vivo, at periodontal disease sites, the procoagulant activity of Rgps is apparently negated by the fibrinogen degradation carried out by Kgp [Scott et aI., 1993; Imamura et aI., 1995a, b], which contributes to a bleeding tendency, a hallmark of the disease, which correlates positively with the presence of P gingivalis at discrete periodontal pockets. Collectively, the interaction of gingipains with the coagulation cascade leads to local, uncontrolled release ofthrombin, an enzyme with a multitude of diverse biological activities, including the stimulation of prostaglandin, IL-l and platelet-activating factor release by endothelial cells and macrophages. These mediators are considered predominant factors in the tissue destruction process in periodontal disease. Another trademark of periodontitis is the increased flow of gingival fluid from periodontal pockets. This symptom can be directly associated with the unrivalled (compared to other bacterial proteases) ability of gingipains to release bradykinin. Physiologically, this potent mediator is released from highmolecular-weight kininogen by plasma kallikrein, which in turn is generated from prokallikrein by activated Hageman factor (factor XIIa). Rgps shortcut this cascade by activation of plasma prekallikrein, with kinetics, which are

Bacterial Peptidases

153

better than those observed in prekallikrein activation by factor XIIa [Imamura et aI., 1994]. In addition, Rgps working in concert with Kgp, can release bradykinin directly from high-molecular-weight kininogen [Imamura et aI., 1995a, b]. Bradykinin exerts powerful biological activities and is responsible for pain and local extravasation at the site of infection/inflammation leading to edema, which underlies the mechanism of generation of gingival crevicular fluid. The main targets for gingipains amongst factors of the complement cascade seem to be the proteins C3 and C5, but the mode of action on these factors is different. While C3 is destroyed, thus disabling the bactericidal and opsonizing ability of activated complement, the functional chemoattractant, C5a, is released from C5 by the action of the gingipains [Wingrove et aI., 1992; Discipio et aI., 1996]. In addition, gingipains can enhance the chemotactic activity ofIL-8 [Mikolajczyk-Pawlinska et aI., 1998]. Cumulatively, this gingipain-mediated generation of potent chemoattractants may lead to excessive neutrophil accumulation at periodontal sites, another clinical sign of active disease. A large set of cell surface proteins and receptors, including the LPS receptor (CD14) [Sugawara et aI., 2000; Tada et aI., 2002], the C5a receptor (CD58) [Jagels et aI., 1996], the IL-6 receptor (lL-6R) [Oleksy et aI., 2002], and ICAM-I [Tada et aI., 2003] are targeted by the gingipains. Although the cleavage of these proteins may significantly contribute to P. gingivalis-induced pathological changes in the periodontium, activation of protease-activated receptors (PARs) deserves special emphasis. PARs mediate cellular responses to a variety of extracellular serine peptidases [Ossovskaya and Bunnett, 2004]. The four known PARs constitute a subgroup of the family of seven-transmembrane domain G protein-coupled receptors and activate intracellular signaling pathways typical for this family of receptors. Activation of PARs involves proteolytic cleavage of the extracellular domain, resulting in formation of a new amino-terminus, which acts as a tethered ligand. PAR-I, PAR-3, and PAR-4 are relatively selective for activation by thrombin whereas PAR-2 is activated by a variety of proteases, including trypsin and tryptase [Gabazza et al., 2004]. Rgps specifically activate intracellular signaling pathways through cleavage ofPAR-2 on neutrophils [Lourbakos et aI., 1998], PAR-I and PAR-4 on platelets [Lourbakos et aI., 2001b], and PAR-I and PAR-2 on human oral epithelial cells [Lourbakos et aI., 200 Ia] with efficiency matching that for the endogenous agonists. Collectively, hijacking of the PAR-dependent signaling pathways illustrates the ability of the gingipains to carry out functional mimicry, which contributes to potentiation of local inflammatory responses and can be directly linked to bone resorption, the most profound clinical sign of advanced periodontal disease.

Potempa/Pike

154

The list of proteins cleaved by gingipains discussed above is far from complete. A more complete set includes P gingivalis extracellular proteins, as well as many other host proteins, such as hemoglobin and heme/iron-binding proteins, cytokines, bactericidal peptides, host peptidase inhibitors, proteins of the extracellular matrix, latent matrix metalloproteinases, and epithelial junctional proteins. The significance of these protein cleavages for periodontal disease pathogenicity is often speculative, but there is no doubt that gingipains carry out an extremely diverse set of interactions with the host. Consistently, strains with the gingipain genes disabled by genetic manipulation have severely decreased virulence [O'Brien-Simpson et aI., 200 I] and the pathogeneicity of P gingivalis can be supressed in vivo by gingipain-specific inhibitors [Curtis et aI., 2002]. Finally, immunization with the gingipains as antigens has protective effects, as observed in animal models of P gingivalis infection [Gibson and Genco, 200 I; Gibson et aI., 2004; Rajapakse et al., 2002].

Clan CE This clan contains five families recognized thus far, three are found exclusively in viruses, one is unique for bacteria (family C55) and one is widespread among cellular organisms, except the archae (family C48). The archetypal enzyme of clan CE is the cysteine peptidase from adenovirus, adenain. Although adenain has a unique scaffold not seen in cysteine peptidases outside clan CE, the active site contains a Cys-His-Glu triplet and an oxyanion hole in an arrangement similar to that in papain [McGrath et aI., 2003; Ding et aI., 1996]. In this respect, the CE clan peptidases represent a powerful example of convergent evolution at the molecular level. Family C48: The Vip] Endopeptidase Family In eukaryotic cells, the modification of proteins by a small ubiquitin-like modifier (SUMO) plays an important role in the function, compartmentalization, and stability of target proteins, contributing to the regulation of diverse processes [Muller et al., 2004; Melchior et al., 2003]. The covalent modification of proteins by SUMO-l is reversible and is mediated by SUMO-specific proteases. These proteases are ubiquitous in eukaryota and are thought to have a dual function. They are responsible firstly for the initial processing of SUMO-I by cleavage ofthe precursor peptide at the carboxyl-terminus of the protein, and secondly for the subsequent processing and cleavage of high molecular weight SUMO-l conjugates, releasing SUMO-I and reducing the conjugation status of the target proteins. Homologues of these peptidases have thus far only been found in a few gram-negative bacteria, including Bradyrhizobium japonicum, Chlamydia muridarum, Chlamydia trachomatis, Mesorhizobium loti, P syringae and X campestris. In the genomes of these organisms, representing animal and

Bacterial Peptidases

155

plant pathogens and plant symbionts, up to 3 genes encoding potentially functional SUMO-specific peptidases are present, but their role in symbiosis or virulence has not been established. However, takjng into account the importance of SUMO conjugation for the functioning of eukaryotic cells [Yeh et aI., 2000], it is tempting to speculate that bacterial homologues of SUMO-specific peptidases are also active inside the host cell, subverting its function to benefit the pathogen, as in the case of the YopJ peptidases described below. Family C55: The YopJ Peptidase Family It is fascinating to note that amongst the type III secretion effectors, human and plant pathogens, as well as plant symbionts, have evolved two conserved families of cysteine peptidases with completely different folds. Both families mimic the proteolytic activity of eukaryotic proteins that are essential for the normal maintenance of host signaling. Members of the YopT family discussed earlier have a typical papain-like fold which has been crafted by pathogen evolution to yjeld a new, specific role in bacterial pathogenjcity. The YopJ family described here apparently evolved using the scaffold of SUMO-specific peptidases (see above). Regardless of their differences in structure and specificity, both groups of enzymes target a limited number of intracellular substrates, specific cleavage of which subdues the host reaction to benefit the invading pathogen. YopJ, one of the effector molecules injected into the host cell by Y pestis was the first protein in this family recognized as a peptidase, based on a comparison ofthe predicted secondary structure ofYopJ to that of the known structure of the adenovirus cysteine peptidase, willch revealed significant similarity between these two proteins [Orth et aI., 2000]. Indeed, the intact catalytic dyad of Cys-His is absolutely necessary for YopJ to exert biological activity in the host eukaryotic cell. Also, the ability of the YopJ homologue, AvrBsT (the effector molecule secreted via the type III pathway by X. campestris pathovar campestris), to trigger the hypersensitive response in plants, was shown to be dependent on the proteolytic activity of AvrBsT. In the case ofYopJ, the activity was exerted by cleaving SUMO-l-conj ugated proteins. Now, it has become clear that plant homologues ofYopJ are also cysteine peptidases with SUMO substrate specificity, since it was shown that XopD, an X. campestris pathovar vesicatoria type III effector injected into plant cells, translocated to subnuclear foci and hydrolyzed SUMO-conjugated proteins in vivo [Hotson et aI., 2003]. Tills indicates that SUMO protein deconjugation is a common strategy utilized by animal and plant pathogens to alter signal transduction. The SUMO-dependant pathway of intracellular signaling is very ancient and evolutionarily conserved in eukaryotic cells. So is its sensitivity to proteolytic interference by YopJ, which cleaves SUMO-conjugated proteins in yeast, resulting in a blockage of

Potempa/Pike

156

the mitogen-activated protein kinase (MAPK) kinase-dependent pathway of signaling [Yoon et aI., 2003]. The cleavage of SUMO conjugates in mammalian cells by Yersinia YopJ peptidase also blocks MAPK kinase [Collier-Hyams et aI., 2002] paralyzing both the innate and adaptive immune responses. There are, however, some differences between the function of different YopJ peptidases, which apparently reflects adaptation to the specific lifestyle of a given pathogen. An AvrA protein from common, mild enteropathogen of humans, S. enterica serovar typhimurium, although 86% similar in amino acid sequence to YopJ, only inhibits NF-KB signaling and augments apoptosis in human epithelial cells, giving rise to speculation that AvrA may limit virulence in vertebrates in a manner analogous to the avirulence factors in plant [CollierHyams et aI., 2002]. The lack of an avrA allele in strains of Salmonella typhi and Salmonella paratyphi [Prager et aI., 2000], which evade epithelial defenses and results in severe systemic diseases seems to support this hypothesis. In summary, in the case of animal pathogens, SUMO protein deconjugation interferes with the innate immune response by blocking cytokine production and inducing apoptosis in the infected cells. The infected host cell cannot respond to invaders because YopJ-like peptidases disrupt an essential posttranslational modification that is required for activation of mammalian MAPK and NF-KB pathways [Orth, 2002].

Clan CF The crystal structures of two peptidases from this clan are known and they are clearly unique. As yet, only one family was distinguished (family C 15). Family C15: The Pyroglutamyl-Peptidase 1 Family Pyroglutamyl-peptidases remove the amino terminal pyroglutamate (pGlu) residue from specific pyroglutamyl substrates [Cummins and O'Connor, 1998]. To date, three distinct forms of this enzyme have been identified, but only type I pyroglutamyl-peptidase is a cysteine peptidase with a unique fold. The active enzyme is apparently a homotetramer [Odagaki et aI., 1999]. Both in mammals and prokaryotes, it is located in the cytoplasm and displays a broad pyroglutamyl substrate specificity. Genes encoding pyroglutamyl-peptidase I occur in several, mostly gram-positive bacterial species, but there are no reports that this enzyme activity may be related to virulence.

Metallopeptidases

Metallopeptidases are hydrolases in which the nucleophilic attack on a peptide bond is carried out by a water molecule activated by a divalent metal

Bacterial Peptidases

157

cation, which is usually zinc, but examples where cobalt, manganese or nickel are used have been reported. The metal ion is usually immobilized by three amino acid ligands, His, Glu, or Asp. In addition to the metal ligands, at least one other residue is involved in catalytic hydrolysis of the peptide bond exercising the functions of a general base in catalytic solvent polarization. In many cases this residue is a glutamate. At present the MEROPS database allocates metallopeptidases to 15 clans recognized by the type and number of metal ions required for catalysis and, within these broad groups, by the sequential arrangement of the metal ligands and the catalytic residue. Within clans, separate families are distinguished based on structural similarities. The most divergent and densely populated clan is MA featuring the zincins, in which the water nucleophile is bound by a single zinc ion ligated to two His residues in a sequential motif of His-Glu-Xaa-Xaa-His, in which Glu is the general base and Xaa stands for any amino acid. Depending on the third Zn ligand, which is either a Glu or His/Asp located downstream of the Zn-binding motif, clan MA is divided into two subclans, MA(E) and MA(M) [Gomjs-Ruth, 2003], respectively. These subclans putatively represent separate evolutionary lines of metallopeptidases after a very ancient divergence within clan MA. Also, peptidases grouped into clan MM utilize the His-Glu-Xaa-XaaHis motif and use an Asp residue to ligate zinc, but they are structurally unrelated to clan MA enzymes. The other well-defined and characteristic sequential motifs involved in zinc chela60n include His-Xaa-Xaa-Glu and His (clan MC), His-Xaa-Xaa-Glu-His and Glu (clan ME), His-Xaa-Glu-Xaa-His with the third ligand unidentified (clan MK), His-Xaa-Xaa-Xaa-Asp and His-Xaa-His (clan (clan MP). MO) and hゥウMs・イ pッHxセIMaウー In contrast to the limited occurrence of aspartic and cysteine peptidases amongst bacteria, metallopeptidases are widespread and they have representatives in 50 out of the 52 distinguished families of this class of enzymes. Even more interestingly, three metallopeptidases, including the FtsH protease [clan MA(E), family M41], methionyl amjnopeptidase (clan MG, family M24), and homologues of sialoglycoprotease from Mannheimia (Pasteurella) haemolytica (Clan MK, family M22) are the only peptidases of any catalytic class which are absolutely conserved among bacterial species. Apparently, this trio features essential house-keeping enzymes and, therefore, a perfect target for the development of inhibitors, which, by blocking the activity of these peptidases, should arrest or kill most bacteria. Methionyl aminopeptidase I is an especially attractive target since the reaction it catalyzes, i.e. removal of the formylated amjnoterminal methionine residue from newly synthesized polypeptide chains, is unique to bacteria. Therefore, one would expect that specific inhibitor of the methionyl aminopeptidase should exert no side effects on eukaryotic organisms, thus resembling the action of classical antibiotics. Unfortunately, however, the

Potempa/Pike

158

mammalian homologues of methionyl aminopeptidase are also susceptible to bacterial enzyme inhibitors. Collectively, the promise of effective new drugs and the obstacles with regard to cross-reactivity has fuelled intense interest in the detailed investigation of this family of peptidases, which are of known tertiary structure, have a characterized mechanism of catalysis and are subject to inhibition by an array of different compounds [Bradshaw et aI., 1998; Bazan et aI., 1994; Douangamath et aI., 2004; Oefner et aI., 2003; Hu et aI., 2004; Towbin et aI., 2003; Copik et aI., 2003; Klein et aI., 2003; Li et aI., 2004]. Using the FtsH protease as a target to fight bacterial infection seems to be an even more challenging task than targeting the methionyl aminopeptidase 1. FtsH is a member of the AAA superfamily (ATPases associated with diverse cellular activities), which includes proteins involved in a variety of cellular processes characterized by conserved regions which include an ATP-binding site and a metal10peptidase domain. These ATP-dependent proteases mediate the degradation of membrane proteins in bacteria, mitochondria and chloroplasts. They combine proteolytic and chaperone-like activities and thus form a membrane-integrated quality control system [Langer, 2000]. In bacteria, the FtsH peptidase is anchored to the cytoplasmic membrane with the catalytic domains exposed to the cytoplasm. In addition to being involved in quality control ofintegral membrane proteins, FtsH peptidase is involved in the posttranslational control of the activity of a variety of important transcription factors [Schumann, 1999]. In this way, FtsH peptidase is involved in the regulation of the stress response together with other chaperones with proteolytic activity, including serine peptidases such as ClpXp, ClpAP, HslUV and Lon [Hengge and Bukau, 2003; Wong and Houry, 2004]. However, unlike the serine peptidase chaperones, FtsH has never been implicated as an agent contributing to pathogenic fitness of a pathogen until recently, when it was shown that as. aureus fisH mutant was attenuated in a murine skin lesion model of pathogenicity [Lithgow et aI., 2004]. The biological function of the sialoglycopeptidase in M. (Pasteurella) haemolytica has been investigated in some detail. The 35-kD enzyme isolated from the culture supernatant of this bacterium is active at neutral pH and is remarkably specific for O-sialoglycoproteins. It cleaves hwnan erythrocyte glycophorin A, which is O-glycosylated at several positions, with a major site of cleavage at Arg3I-Asp32, but does not cleave N-glycosylated proteins or nonglycosylated proteins [Abdullah et aI., 1992]. The importance of the enzyme in the pathogenicity of bovine pneumonic pasteurellosis (shipment fever) caused by M. (Pasteurella) haemolytica is not clear, although the enzyme may interfere with cell-cell adhesion or with cytokine receptor binding through the cleavage of the cell surface O-sialoglycoproteins [Sutherland et aI., 1992] during the development of the host immune response in the cattle lung. Also, the sialoglycopeptidase-mediated enhanced adhesion to bovine platelets may

Bacterial Peptidases

159

initiate platelet aggregation and fibrin formation in alveolar tissue in pneumonic pasteurellosis [Nyarko et aI., 1998]. Genes encoding potentially active homologues of the sialoglycopeptidase are conserved across all cellular forms of life, but their biological function is still a puzzle. The essentiality nature of this gene for some bacteria indicates that the enzyme has a very important biological function, but either we do not know its physiological substrate(s) or the protein carries out a function unrelated to proteolytic activity. At least in the case Schizosaccharomyces pombe the sialoglycopeptidase homologue has been shown to be involved in pro-protein processing [Ladds and Davey, 2000]. The large number of bacterial metaHopeptidases excludes the possibility of a systematic description of each family of these peptidases in the context of their involvement in pathogenicity. It is interesting to note that a relatively large number of peptidase families in clans MA(E) (7 out of 16) and MA(M) (6 out of 12) have no counterparts in any other cellular form of life outside the (archae) bacterial kingdom. In addition to peptidases, which are strongly implicated as virulence factors, only members of families specific for bacteria are discussed below in more detail.

Family M4: Thennolysin Family Thermolysin, an extracellular metallopeptidase isolated from Bacillus thermoproteolyticus, constitutes an archetype, not only of this family, but also for bacterial metallopeptidases in general. Enzymes homologous to thermolysin are expressed by several pathogens, including L. monocytogenes, S. epidermidis, S. aureus, Enterococcus faecalis, c. perfringens, Helicobacte pylori, P aeruginosa and V cholerae. Their involvement in pathogenicity is generally related to the broad substrate specificity of these peptidases, which can attack several physiologically important host proteins. A significant amount of data has been generated regarding the destructive function of pseudolysin from P aenlginosa, an enzyme known for its strong elastinolytic activity [Wretlind and Wadstrom, 1977; Galloway, 1991]. This peptidase, also referred to as P aeruginosa elastase, exerts its destructive action by direct degradation of several connective tissue proteins [Kessler et aI., 1977; Heck et aI., 1986; Galloway, 1991] and, indirectly, by inactivation of host proteinase inhibitors, including ai-antitrypsin [Morihara et aI., 1979]. Through its fibrinogenolytic and fibrinolytic activities, the elastase may disturb homeostasis and induce changes in the structure of the vascular wall, causing leakage of the plasma component, including cells into the extravascular tissue. This activity can potentially induce a hemorrhagic tendency and damage of infected tissue [Komori et aI., 2001]. In lungs, the enzyme may degrade surfactant proteins SP-A and SP-D, which have important roles in the innate immune response.

Potempa/Pike

160

This mechanism significantly contributes to the virulence mechanism in the pathogenesis of chronic P aeruginosa infection [Mariencheck et aI., 2003]. This data correlate well with the observation suggesting that the P aeruginosa elastase is a potent inflammatory factor in a mouse model of diffuse panbronchiolitis [Yanagihara et aI., 2003] and that the control of elastase release by P aeruginosa may be beneficial for patients with diffuse panbronchiolitis. Also, pseudolysin seems to play an essential role in the initiation and/or maintenance of a corneal infection [Hobden, 2002]. The role of pseudo lysin orthologues in other pathogenic bacteria is less well understood and requires further investigation. Nevertheless, aureolysin from S. aureus has been shown to contribute to connective tissue degradation by host peptidases through inactivation of host proteinase inhibitors [Potempa et aI., 1986, 1991]. It may also assist in S. aureus dissemination by degradation of bacterial adhesins [McAleese et aI., 2001]. A similar function is suggested for the hemagglutinin/peptidase of V. cholerae, which may be responsible for the detachment of these bacteria from cells through digestion of several putative adhesion receptors [Finkelstein et aI., 1992]. On the other hand, the L. pneumophila Msp protease can significantly suppress antibacterial human phagocyte responses and contribute to the pathogenesis of Legionnaire's illsease [Sahney et aI., 2001]. A totally different mechanism seems to be utilized by the gelatinase (GelE) secreted by E. faecalis. This enzyme, which is also termed coccolysin, is implicated as a virulence factor by both epidemiological data and animal model studies and can apparently contribute to the dissemination of E. faecalis by fibrin degradation [Waters et aI., 2003]. It is also possible that some of the manifestations of inflammatory conditions in the presence of E. faecalis are related to coccolysincatalyzed inactivation of endothelin [Makinen and Makinen, 1994].

Family M6: Immune Inhibitor A Family The name ofthis family, also known as thuringilysin family and belonging to the metzincin clan (MA(M» [Gomis-Ruth, 2003], refers to the ability of proteins initially isolated from Bacillus thuringiensis to inactivate the antibacterial activity of insect hemolymph [Edlund et aI., 1976]. It is now known that this protein is a metallopeptidase, exerting its insecticidal activity by proteolytic degradation of attacins and cecropins, two classes of antibacterial proteins in insects, and thus kills insect larvae [Dalhammar and Steiner, 1984; Lovgren et al., 1990]. This unique property contributes to the use of B. thuringiensis in biological pest control. Fortunately, this kind of peptidase, which is very effective in disabling the most important weapon ofthe host innate defense, is limited to insect pathogens. Nevertheless, several bacterial peptidases of different catalytic classes have been described to be able to inactivate human antibacterial peptides, once again indicating the importance of this activity in bacterial pathogenesis.

Bacterial Peptidases

161

Family M9: Microbial Collagenase By virtue of being able to degrade collagen, one of the major proteinaceous constituents of the connective tissue and extracellular matrix, bacterial peptidases with this activity are by default recognized as virulence factors [Harrington, 1996]. The members of this family are common among Clostridium spp., Bacillus spp., and Vibrio spp. Despite the potential ability to inflict extensive tissue damage and facilitate spreading of infection, the precise role of microbial collagenases in pathogenicity remains unclear. Family MIO This family is divided into two subfamilies in MEROPS, though according to somewhat dubious criteria. Both belong to the metzincin clan [Gomis-Ruth, 2003], as well as those ofthe -equally cryptically subdivided- family 12. Subfamily lOA encompasses predominantly eukaryotic MMPs. Probable orthologues have been identified in the genomes of archaebacteria (Methanosarcina acetivorans, Methanosarcina mazei Gol, Methanosarcina barkeri), uncultured crenarchaeote, and bacteria (Bacillus anthracis, Listeria innocua, L. monocytogenes, Leptospira inten'ogans, and S. pneumoniae). In the latter cases, function as putative virulence factors or housekeeping enzymes remains to be assessed. According to MEROPS, subfamily lOA would further encompass a secreted 20-kD metallopeptidase toxin, B.fragilis toxin (BFT). The toxin also known as fragilysin is considered an important factor in the pathogenicity of infections with enterotoxigenic B. fragilis (ETBF), a recently identified enteric pathogen of children and adults. Fragilysin can directly damage human colonic mucosa [Riegler et aI., 1999]. This effect is apparently dependent on cleavage ofE-cadherin, the primary protein of the zonula adherens, leading to modification ofepithelial cell morphology in vitro and resulting in increased fluid secretion into the intestine, which is clinically manifested as diarrhea [Wu et al., 1998; Sears, 2001]. Also, fragilysin contributes to intestinal mucosal inflammation by stimulation of the expression of the neutrophil chemoattractant cytokine, IL-8 [Sanfilippo et aI., 2000]. According to another classification, fragilysin, together with three paralogues and an orthologue in the photosynthetic cyanobacteriwn Nostoc punctiforme, would constitute an independent family within the metzincins, though structurally probably related to MMPs [Gomis-Ruth,2003]. Only bacterial peptidases are grouped in subfamily lOB, which are exemplified by the major metalloproteinase secreted by Serratia marcescens, termed serralysin. The other members of the subfamily include aeruginolysin, an alkaline protease from P aeruginosa, mirabilysin (ZapA protease) from Proteus mirabilis, and several peptidases from Erwinia spp. Aeruginolysin seems to play a major role in the pathogenesis of eye infections by enhancing P aeruginosa attachment to corneal epithelium [Pillar et aI., 2000] and is a target for vaccine

Potempa/Pike

162

development, and chemotherapy for bacterial eye infections. On the other hand, mirabilysin is considered to be an important virulence factor because it degrades host immunoglobulins, contributing to immune evasion during urinary tract infection [Walker et aI., 1999; Almogren et aI., 2003].

Family M26: /gAl-Specific Peptidase Many of the important mucosal bacterial pathogens, including Haemophilus influenzae, Neisseria gonorrhoeae, Neisseria meningitides, S. pneumoniae and successful members of the human resident flora, such as Streptococcus mitis, Streptococcus oralis, and Streptococcus sanguinis, have developed peptidases exclusively specific for cleavage at the hinge region of IgA 1. These peptidases apparently belong to three catalytic classes, but only enzymes belonging to the serine (family S6) and metal10peptidase (family M26) classes have been thoroughly characterized. The IgA I-metallopeptidases are produced by Streptococcus spp., with a significant exception being GAS (s. pyogenes), while Haemophilus and Neisseria spp. produce serine-type IgA peptidases. Taken together, these peptidases are a striking example of convergent evolution to the same function by bacterial virulence factors [Kilian et aI., 1996]. All these enzymes cleave peptide bonds at a PI proline residue within the hinge region of IgA I, separating the antigen-binding Fab fragment from the Fc fragment. This mode of cleavage, which removes the Fc effector domain of the IgAI molecule, not only eliminates the protective effect of the immunoglobulins, but can also serve to camouflage the bacteria with Fab fragments, which mask the epitopes recognized by intact, functional antibodies. Despite this narrow specificity, which is precisely aimed to not only disable the effector molecules of host immune system and to take advantage of them, the exact role of these enzymes in bacterial pathogenesis is still unclear. This is due to the lack of an appropriate animal model to test the contribution of these enzymes to pathogenicity, since they only cleave human, gorilla or chimpanzee IgA I molecules [Reinholdt and Kilian, 1991]. In the context of convergent evolution it is worth mentioning the IgA specific metallopeptidase produced by Clostridium ramosum here (family M64) [Kosowska et aI., 2002]. This enzyme has specificity for cleavage of both IgAI and IgA2 molecules, which is a clear adaptation to the commensal lifestyle in the human gut, where both IgA isotypes are abundant. Family M27: Tentoxilysin Neurotoxins produced by several serotypes of Clostridium botulinum (BoNT type A-G) and Clostridium tetanum (TeNT) are the most potent natural toxins known to date. The toxins exert their biological effects at subfemtomolar

Bacterial Peptidases

163

concentrations and they are released into the environment upon bacterial lysis as a single polypeptide chain of 150 kD. Proteolytic cleavage executed by host peptidases generates a two-chain, mature, active neurotoxin composed of a heavy chain (100 kD) and a light chain (50 kD) held together by a single disulfide bridge. The heavy chain is responsible for the specific binding of the toxin to presynaptic membranes and the translocation of the light chain into the nemon. The light chain is a very specific metallopeptidase with activity limited to a small subset of proteins, including VAMP/synaptobrevin, SNAP-25 and syntaxin, which play key roles in synaptic signal transduction [Schiavo et aI., I992a, b; Montecucco and Schiavo, 1994]. Cleavage of these proteins directly leads to the clinical manifestations of tetanus and botulism. Cumulatively, tentoxilysins represent a very interesting example of the development of extremely specific and potent virulence factors. Fortunately, their occurrence is limited to a few Clostridium spp.

Family M34: Anthrax Lethal Factor The anthrax toxin is one of the most lethal natmal toxins. It is produced by Bacillus anthracis and spores of these bacteria are the active component of the most deadly bioweapon developed by mankind. The toxin is composed of three proteins, including protective antigen (PA), edema factor (EF) and lethal factor (LF). PA binds to specific cell smface receptors and, upon proteolytic activation by cell membrane-associated furin-like host peptidases, forms a membrane channel through which EF and LF enter the cell. LF is a unique multidomain metallopeptidase with a very narrow specificity to cleave the amino-terminus of mitogen-activated kinase kinases 1 and 2 (MMPKKI and MMPKK2). The cleavage inactivates the signal transduction pathway dependent on these kinases. This signaling pathway plays a fundamental role in the overall intracellular signaling network, thus the overall signaling in the cell is compromised. Family M56: BlaRl Peptidase (S. ameus) The BlaRl peptidase from S. aureus is a metallopeptidase which cleaves a repressor (Blal) of the synthesis of the 13-laclamase enzyme BlaZ by this bacterium [Hackbarth and Chambers, 1993]. Thus, this peptidase controls antibiotic resistance by controlling the production of the 13-lactamase. The BlaR I peptidase orthologue, Mec Rl, only found in methicillin resistant S. aureus (MRSA), controls the formation ofthe penicillin-binding protein 2a (PBP 2a) and thereby controls the resistance of the bacterium to methicillin [Hackbarth and Chambers, 1993; Brakstad and MaeJand, 1997]. The BlaRl molecule consists of two domains, an extracellular penicillin-binding domain and an integral-membrane zinc metallopeptidase domain [Zhang et aI., 200 I]. Upon penicillin binding, the BlaR I peptidase autoactivates, then cleaves the repressor of 13-lactamase

Potempa/Pike

164

synthesis, providing an interesting 'signal transduction' system which mediates this antibiotic resistance in the highly pathogenic staphylococcus species. Family M66: StcE Protease The StcE metallopeptidase, member of the cholorerilysins within the metzincin clan MA(M) [Gomis-Riith, 2003] is produced by the enterohemorrhagic 0157:H7 strain of E. coli, which causes diarrhea, hemorrhagic colitis, and the hemolytic uremic syndrome, specifically cleaves C I inhibitor (also known as CI esterase inhibitor). The peptidase is quite specific for CI inhibitor and does not appear to cleave other proteins, although it has been shown to cause aggregation of cultured T cells, the significance of which is not completely understood [Lathem et aI., 2002]. C I inhibitor is known to control potent proinflammatory and procoagulant enzymes, and thus its inactivation by the bacterial peptidase is likely to cause proinflammatory effects which may be consistent with the disease outcomes caused by this strain of E. coli. Further experiments will be required to elucidate how critical this enzyme is to pathogenesis by this strain of the bacterium. Family M73: Camelysin Camelysin (casein-cleaving metalloprotease) is found on the surface of B. cereus, whose genome encodes a total of four paralogues. Possible orthologues have been identified in the genomes of Oceanobacillus iheyensis (five sequences) and B. anthracis (two sequences). Single sequences are further found in B. thuringiensis, B. subtilis, and Bacillus halodurans (Gomis-Riith; personal communication). This bacterium is known to cause food poisoning and nosocomial diseases. Camelysins do not have a sequence consistent with metalloproteases, but the enzyme is active against a broad range ofproteinaceous substrates and mass spectrometry analyses strongly indicate the association of a zinc ion with each enzyme molecule. Disruption of the gene for the enzyme causes a marked loss in the proteolytic activity of membranes from the bacterium and it is possible that the enzymatic activity plays a role in the pathogenic activity 0 the organism, although this remains to be fimlly established [Grass et aI., 2004].

Serine Peptidases

Peptidases which utilize a serine residue as the main catalytic residue are the biggest group ofpeptidases, making up 35% of the total peptidases listed in MEROPS. The serine peptidases are widespread across all organisms and are divided into 10 clans on the MEROPS database [SB, SC, SE, SF, SH, SJ, SK, SP, SR and S- (the last contains currently unassigned peptidases)]. Bacterial

Bacterial Peptidases

165

proteases are present in all of these clans, except SH, SP and SR, which will therefore not be considered any further here. By definition, this catalytic class contains a serine residue acting as the nucleophile during catalysis. Usually (as applies to enzymes in clans SB, SC and SK) the catalytic Ser residue combines with His and Asp residues to form the classical catalytic triad exemplified by chymotrypsin, the archetypal enzyme of the serine protease class. Variations on this do exist, for instance enzymes in clans SE and SJ use a SerlLys dyad to accomplish catalysis, while those in SF use either a Ser/Lys or a Ser/His dyad. There are over 60 families represented within the serine-type catalytic class and many of these are subdivided into subfamilies. The sheer number of proteases in this catalytic class which are found in bacteria defies their being mentioned in any representative manner here. Thus the most interesting or wellcharacterized examples with direct relevance in pathogenicity were selected for presentation here. Family SiB The glutamyl endopeptidase I, better known as endoproteinase GluC or the V8 protease from S. aureus, is a member of the SIB famjly. The roles of tills enzyme are somewhat related to pathogenicity (see section Family C47: The Staphopain Famjly above), but this enzyme is better known for its widespread biotechnological use as a specific protease in sequencing applications. Its structure has recently been solved [Prasad et aI., 2004]. This family also contains the Spl peptidases, which have recently been identified as a new operon which is positively controlled by the Agr vjrulence regulator, indicating a possible role in pathogenesis by S. aureus [Reed et aI., 2001]. Family SiC An interesting group of peptidases is formed by members ofthe SIC family, which is required for growth at high temperatures by a number of organisms, such as E. coli. Some of these enzymes, generically termed protease Do (also referred to as DegP or HtrA), have been characterized as being associated with the virulence of S. enterica serovar typhimurium, Yersinia enterocolitica and S. pyogenes. DegP from E. coli has a fascinating dual function of acting as a chaperone and a peptidase, depending on the temperature of the environment. In the chaperone phase, a hydrophobic patch of amino acids plays the presumptive role of binding unfolded proteins and mediating their refolding. During chaperone operation, the active site for the peptidase is 'walled off', preventing substrate binding and catalysis. A change in the envjronmental conditions triggers the opening of the active site to substrates and allows catalysis. This fascinating mechamsm allows the peptidase to process many different proteins needed for pathogenesis by the bacteria.

Potempa/Pike

166

Family SiD The family is entirely composed of the endoproteinase lysC and endoproteinase Arg-C, which have applications in the sequencing of proteins due to their high specificity for lysine and arginine amino acids at the cleavage point, respectively. An endoproteinase Arg-C orthologue from P aeruginosa is thought to act as a virulence factor in corneal infections by this bacterium [Engel et aI., 1998]. Family S6 The IgA I-specific serine endopeptidases which are found in Neisseria spp. and some Haemophilus spp. are typical members of the S6 family. In N. gonorrhoeae, the enzyme has been postulated to playa role in evading the host immune response by specifically cleaving IgA 1 [Vitovski et aI., 1999]. It has been suggested that the enzyme plays a role in bacterial invasion of host cells [Lin et aI., 1997]. However, whether the IgA I-specific serine endopeptidase is a crucial virulence factor has yet to be determined [Johannsen et aI., 1999]. Family S8A This group of serine proteases contain enzymes generally referred to as subtilisin-like enzymes, named after the archetypal enzyme of the group. The family contains a large number of enzymes, most likely second only to family SIA which contains the mammalian chymotrypsin-like enzymes. The subtilisins and chymotrypsin-like enzymes are examples of convergent evolution, arriving at the same function and catalytic groups, but grafted onto very different scaffolds. Perhaps the best-characterized virulence factor ofthis family is the C5a peptidase from group A and group B Streptococci, exemplified by the enzyme from S. pyogenes. As the name suggests, this enzyme cleaves the C5a component of complement, destroying its ability to act as a chemotaxin for polymorphonuclear leukocytes [Hill et aI., 1988]. Recent studies suggest that this enzyme is also able to bind to fibronectin, which may be important in the binding and invasion of host cells by group B streptococci [Beckmann et aI., 2002; Cheng et aI., 2002b]. Recently, much effort has been invested into the development of C5a peptidasebased vaccines for the treatment of group A and B streptococcal infections [Shet et aI., 2003; Cheng et aI., 2002a]. Family S9B Members of the family S9B are generally dipeptidyl peptidases, which cleave two amino acids at a time from the terminii of proteins. The bacterial peptidases in this subfamily are exemplified by the dipeptidyl aminopeptidase

Bacterial Peptidases

167

1\1, from organisms such as P gingivalis [Banbula et aI., 2000; Kumagai et aI., 2000]. The enzyme is apparently important for the virulence of P gingivalis, since bacteria lacking the protease or with a mutation in the catalytic domain have attenuated virulence [Kumagai et aI., 2003].

Family 514 and 516 The S14 family is primarily composed of the endopeptidase Clp enzymes, originally discovered and characterized in E. coli. Endopeptidase Clp enzymes are rather similar to Lon proteases (S 16 family) in that their activity as a peptidase is linked to the hydrolysis of ATP. The enzymes contain an ATP binding and catalysis domain and a distinct peptidase domain [Wang et aI., 1997]. Some studies suggest that these enzymes are the functional equivalents of the proteasome complex found in all mammalian cells, which is crucial for the control of protein turnover in these cells. Interesting support for this hypothesis is provided by a recent study which suggests that the Clp enzyme is important for survival of bacteria which are in the stationary phase [Weichart et aI., 2003]. The catalytic dyad of Lon proteases consists of Ser and Lys. The enzyme is normally induced under stress conditions [Botos et aI., 2004], and animal studies suggest it is highly important S. enterica serovar typhimurium virulence [Takaya et aI., 2003].

Conclusions

As is evidenced by the above review, which is by necessity not absolutely comprehensive, there is a wealth of information about bacterial peptidases. In many instances, however, knowledge is just starting to be accumulated about specific families or enzymes within families. Bacterial peptidases span a tremendous range of mechanisms, and frequently have surprising associations with additional domains which carry out separate functions. This adds a fascinating range to the potential activities of these enzymes. In many cases, the potential for inhibitors of the enzymes to be used as antibacterial agents will continue to drive the active and thriving research in this important field.

References Abdullah KM, Udoh EA, Shewen PE, Mellors A: A neutral glycoprotease of Pasteurella haemolytica A I specifically cleaves O-sialoglycoproteins. Infect Immun 1992;60:56-62. Almogren A, Senior BW, Loomes LM, Kerr MA: Structural and functional consequences of cleavage of human secretory and human serum immunoglobulin A I by proteinases from Proteus mirabilis and Neisseria meningitidis. Infect Immun 2003;71 :3349-3356.

Potempa/Pike

168

Anantha RP, Stone KD, Donnenberg MS: Effects of bfp mutations on biogenesis of functional enteropathogenic Escherichia coli type IV pili. J Bacteriol 2000; 182:2498-2506. Aravind L, Koonin EV: Classification of the caspase-hemoglobinase fold: Detection of new families and implications for the origin of the eukaryotic separins. Proteins 2002;46:355-367. Ashbaugh CD, Warren HB, Carey VJ, Wessels MR: Molecular analysis of the role of the group A streptococcal cysteine protease, hyaluronic acid capsule, and M protein in a murine model of human invasive soft-tissue infection. J Clin Invest 1998; I02:550-560. Ashbaugh CD, Wessels MR: Absence of a cysteine protease effect on bacterial virulence in two murine models of human invasive group A streptococcal infection. Infect Immun 200 1;69:6683-6688. Axtell MJ, Chisholm ST, Dahlbeck 0, Staskawicz BJ: Genetic and molecular evidence that the Pseudomonas syringae type III effector protein AvrRpt2 is a cysteine protease. Mol Microbiol 2003;49: 1537-1546. Axtell MJ, Staskawicz BJ: Initiation of RPS2-specified disease resistance in Arabidopsis is coupled to the AvrRpt2-directed elimination ofRIN4. Cell 2003;1 12:369-377. Aziz RK, Pabst MJ, Jeng A, Kansal R, Low DE, Nizet V, Kotb M: Invasive M ITI group A Streptococcus undergoes a phase-shift in vivo to prevent proteolytic degradation of multiple virulence factors by SpeB. Mol MicrobioI2004;51:123-134. BanbuJa A, Bugno M, Goldstein J, Yen J, Nelson 0, Travis J, Potempa J: Emerging family ofproline-specific peptidases of Porphyromonas gingivalis: Purification and characterization of serine dipeptidyl peptidase, a structural and functional homologue of mammalian prolyl dipeptidyl peptidase IV Infect Immun 2000;68:1176-1182. BarrettAJ, Tolle Dp, Rawlings ND: Managing peptidases in the genomic era. Bioi Chern 2003;384:873-882. Bazan JF, Weaver LH, Roderick SL, Huber R, Matthews BW: Sequence and structure comparison suggest that methionine aminopeptidase, prolidase, aminopeptidase P, and creatinase share a common fold. Proc Natl Acad Sci USA 1994;91 :2473-2477. Beckmann C, Waggoner JD, Harris TO, Tamura GS, Rubens CE: Identification of novel adhesins from group B Streptococci by use of phage display reveals that C5a peptidase mediates fibronectin binding. Infect Immun 2002;70:2869-2876. Berge A, Bjorck L: Streptococcal cysteine proteinase releases biologically active fragments of streptococcal surface proteins. J Bioi Chern 1995;270:9862-9867. Berge A, Kihlberg BM, Sjoholm AG, Bjorck L: Streptococcal protein H forms soluble complementactivating complexes with IgG, but inhibits complement activation by IgG-coated targets. J BioI Chern 1997;272:20774-20781. Bierne H, Garandeau C, Pucciarelli MG, Sabet C, Newton S, Garcia-del Portillo F, Cossart p, Charbit A: Sortase B, a new class of sortase in Listeria monocytogenes. J Bacteriol 2004; 186: 1972-1982. Bjorck L, Akesson P, Bohus M, Trojnar J, Abrahamson M, Olafsson I, Grubb A: Bacterial growth blocked by a synthetic peptide based on the structure of a human proteinase inhibitor. Nature 1989;337:385-386. Botos I, Melnikov EE, Cherry S, Tropea JE, Khalatova AG, Rasulova F, Dauter Z, Maurizi MR, Rotanova TV, Wlodawer A, Gustchina A: The catalytic domain of E. coli Lon protease has a unique fold and a Ser-Lys dyad in the active site. J Bioi Chern 2004;279:8140-8148. Bourgeau G, Lapointe H, Peloquin P, Mayrand 0: Cloning, expression, and sequencing ofa protease gene (tpr) from Porphyromonas gingiva/is W83 in Escherichia coli. Infect Immun 1992;60:3186-3192. Bradshaw RA, Brickey WW, Walker KW: N-terminal processing: The methionine aminopeptidase and N alpha-acetyl transferase families. Trends Biochem Sci 1998;23:263-267. Brakstad OG, Maeland JA: Mechanisms of methicillin resistance in staphylococci. APMIS 1997;105: 264-276. Burns EH Jr, Marciel AM, Musser JM: Activation ofa 66-kilodalton human endothelial cell matrix metalloprotease by Streptococcus pyogenes extracellular cysteine protease. Infect Immun 1996;64: 4744-4750. Calander AM, Jonsson 1M, Kanth A, Arvidsson S, Shaw L, Foster SJ, Tarkowski A: Impact ofstaphylococcal protease expression on the outcome of infectious arthritis. Microbes Infect 2004;6: 202-206. Chan PF, Foster SJ: Role of SarA in virulence determinant production and environmental signal transduction in Staphylococcus aureus. J Bacteriol 1998; 180:6232-6241.

Bacterial Peptidases

169

Chen CY, Luo SC, Kuo CF, Lin YS, Wu JJ, Lin MT, Liu CC, Jeng WY, Chuang WJ: Maturation processing and characterization of streptopain. J Bioi Chern 2003;278: 17336-17343. Chen JM, Rawlings NO, Stevens RA, Barrett AJ: Identification of the active site of legumain links it to caspases, clostripain and gingipains in a new clan of cysteine endopeptidases. FEBS Lett 1998; 441:361-365. Cheng LW, Schneewind 0: Type]]] machines of Gram-negative bacteria: Delivering the goods. Trends MicrobioI2000;8:214-220. Cheng Q, Debol S, Lam H, Eby R, Edwards L, Matsuka Y, Olmsted SB, Cleary PP: Immunization with C5a peptidase or peptidase-type III polysaccharide conjugate vaccines enhances clearance of group B Streptococci from lungs of infected mice. Infect Immun 2002a;70:6409-64 I5. Cheng Q, Stafslien 0, Purushothaman SS, Cleary P: The group B streptococcal C5a peptidase is both a specific protease and an invasin. Infect Immun 2002b;70:2408-2413. Collier-Hyams LS, Zeng H, Sun J, Tomlinson AD, Bao ZQ, Chen H, Madara JL, Orth K, Neish AS: Cutting edge: Salmonella AvrA effector inhibits the key proinflammatory, anti-apoptotic NF-kappa B pathway. J Immunol 2002; 169:2846-2850. Collin M, Olsen A: Generation of a mature streptococcal cysteine proteinase is dependent on cell wallanchored MI protein. Mol Microbiol 2000;36:1306-1318. Collin M, Svensson MD, Sjoholm AG, Jensenius JC, Sjobring U, Olsen A: EndoS and SpeB from Streptococcus pyogenes inhibit immunoglobulin-mediated opsonophagocytosis. Infect Immun 2002;70:6646-6651. Comfort D, Clubb RT: A comparative genome analysis identifies distinct sorting pathways in grampositive bacteria. Infect Immun 2004;72:2710-2722. Connolly KM, Smith BT, Pilpa R, Ilangovan U, Jung ME, Clubb RT: Sortase from Staphylococcus aureus does not contain a thiolate-imidazolium ion pair in its active site. J Bioi Chern 2003;278: 34061-34065. Copik AJ, Swierczek SI, Lowther WT, D'souza VM, Matthews BW, Holz RC: Kinetic and spectroscopic characterization of the H 178A methionyl aminopeptidase from Escherichia coli. Biochemistry 2003;42:6283-6292. Cornelis GR: The Yersinia Ysc-Yop 'type 111' weaponry. Nat Rev Mol Cell Bioi 2002;3:742-452. Cornelis GR, Van Gijsegem F: Assembly and function of type III secretory systems. Annu Rev Microbiol 2000;54:735-774. Coulter SN, Schwan WR, Ng EY, Langhorne MH, Ritchie HD, Westbrock-Wadman S, Hufnagle WO, Folger KR, Bayer AS, Stover CK: Staphylococcus aureus genetic loci impacting growth and survival in multiple infection environments. Mol Microbiol 1998;30:393-404. Cummins PM, O'Connor B: Pyroglutamyl peptidase: An overview of the three known enzymatic forms. Biochim BiophysActa 1998;1429:1-17. Curtis MA, Aduse Opoku J, Rangarajan M, Gallagher A, Sterne JA, Reid CR, Evans HE, Samuelsson B: Attenuation of the virulence of Porphyromonas gingivalis by using a specific synthetic Kgp protease inhibitor. Infect Ill1mun 2002;70:6968-6975. Curtis MA, Kuramitsu HK, Lantz M, Macrina FL, Nakayama K, Potempa J, Reynolds EC, Aduse-Opoku J: Molecular genetics and nomenclature of proteases of Porphyromonas gingivalis. J Periodontal Res 1999;34:464-472. Dalhammar G, Steiner H: Characterization of inhibitor A, a protease from Bacillus thuringiensis which degrades attacins and cecropins, two classes of antibacterial proteins in insects. Eur J Biochem 1984; 139:247-252. Ding J, McGrath WJ, Sweet RM, Mangel WF: Crystal structure of the human adenovirus proteinase with its II amino acid cofactor. EMBO J 1996; 15: 1778-1783. Discipio RG, Daffern PJ, Kawahara M, Pike R, Travis J, Hugli TE, Potempa J: Cleavage of human complement component C5 by cysteine proteinases from Porphyromonas (Bacteroides) gingivalis. Prior oxidation of C5 augments proteinase digestion of C5. Immunology 1996;87:660-667. Donham MC, Heath HE, LeBlanc PA, Sloan GL: Characteristics of extracellular protein production by Staphylococcus simulans biovar staphylolyticus during aerobic and anaerobic growth. J Gen Microbiol 1988; 134:2615-2621. Douangamath A, Dale GE, D' Arcy A, Almstetter M, Eckl R, Frutos-Hoener A, Henkel B, Illgen K, Nerdinger S, Schulz H, MacSweeney A, Thonnann M, Treml A, Pierau S, Wadman S, Oefner C:

Potempa/Pike

170

Crystal structures of Staphylococcus aureus methionine aminopeptidase complexed with keto heterocycle and aminoketone inhibitors reveal the formation of a tetrahedral intermediate. J Med Chern 2004;47: 1325-1328. Drapeau GR: Role ofmetalloprotease in activation of the precursor of staphylococcal protease. J Bacteriol 1978;136:607-613. Dubin G, Chmiel 0, Mak P, Rakwalska M, Rzychon M, Dubin A: Molecular cloning and biochemical characterisation of proteases from Staphylococcus epidermidis. Bioi Chern 2001 ;382: 1575-1582. Dubin G, Krajewski M, Popowicz G, Stec-Niemczyk J, Bochtler M, Potempa J, Dubin A, Holak TA: A novel class of cysteine protease inhibitors: Solution structure of staphostatin A from Staphylococcus aureus. Biochemistry 2003;42: 13449-13456. Dubin G, Stec-Niemczyk J, Dylag T, Silberring J, Dubin A, Potempa J: Characterization ofa highly specific, endogenous inhibitor of cysteine proteases from Staphylococcus epidermidis, a new member of the staphostatin family. Bioi Chern, in press. Edlund T, Siden I, Boman HG: Evidence for two immune inhibitors from Bacillus thuringiensis interfering |セエィ the humoral defense system of satumiid pupae. Infect Immun 1976; 14:934-941. Egile C, d'Hautevilie H, Parsot C, Sansonetti PJ: SopA, the outer membrane protease responsible for polar localization oflcsA in Shigellajlexneri. Mol Microbiol 1997;23:1063-1073. Eichinger A, Beisel HG, Jacob U, Huber R, Medrano FJ, Banbula A, Potempa J, Travis J, Bode W: Crystal structure of gingipain R: An Arg-specific bacterial cysteine proteinase with a caspase-like fold. EMBO J 1999; 18:5453-5462. Eijsink VG, Axelsson L, Diep DB, Havarstein LS, Holo H, Nes IF: Production of class II bacteriocins by lactic acid bacteria; an example of biological warfare and communication. Antonie Van Leeuwenhoek 2002;81 :639-654. Elliott SO: A proteolytic enzyme produced by group A streptococci with special reference to its effect on the type-specific M antigen. J Exp Med 1945;81:573-592. Engel LS, Hill JM, Caballero AR, Green LC, O'Callaghan RJ: Protease IV, a unique extracellular protease and vimlence factor from Pseudomonas aeruginosa. J Bioi Chern 1998;273: 16792-16797. Eriksson A, Norgren M: Cleavage of antigen-bound immunoglobulin G by SpeB contributes to streptococcal persistence in opsonizing blood. Infect Immun 2003;71:211-217. Fenster KM, Parkin KL, Steele JL: Characterization of a thiol-dependent endopeptidase from Lactobacillus helveticus CNRZ32. J Bacteriol 1997; 179:2529-2533. Filipek R, Rzychon M, Oleksy A, Gruca M, Dubin A, Potempa J, Bochtler M: The staphostatin-staphopain complex: A forward binding inhibitor in complex with its target cysteine protease. J BioI Chern 2003;278:40959-40966. Finkelstein RA, Boesman-Finkelstein M, Chang Y, Hase CC: Vibrio cholerae hemagglutinin/protease, colonial variation, virulence, and detachment. Infect Immun 1992;60:472-478. Fischetti VA: Novel method to control pathogenic bacteria on human mucous membranes. Ann NY Acad Sci 2003;987:207-214. Gabazza EC, Taguchi 0, Kamada H, Hayashi T, Adachi Y, Suzuki K: Progress in the understanding of protease-activated receptors. Int J Hematol 2004;79: 117-122. Galloway DR: Pseudomonas aeruginosa elastase and elastolysis revisited: Recent developments. Mol Microbiol 1991 ;5:2315-2321. Gerlach 0, Knoll H, Kohler W, Ozegowski JH, Hribalova V: Isolation and characterization of erythrogenic toxins. V Communication: Identity of erythrogenic toxin type B and streptococcal proteinase precursor. Zentralbl Bakteriol Mikrobiol Hyg 1983;255:221-233. Gibson FC III, Genco CA: Prevention of Porphyromonas gingivalis-induced oral bone loss following immunization with gingipain Rl.lnfect Immun 2001;69:7959-7963. Gibson FC Ill, Gonzalez DA, Wong J, Genco CA: Porphyromonas gingivalis-specific immunoglobulin G prevents P gingivalis-elicited oral bone loss in a murine model. Infect Immun 2004;72: 2408-2411. Gomis-Riith FX: Structural aspects of the metzincin clan ofmetalloendopeptidases. Mol Biotech 2003; 24: 157-20 I. Grass G, Schierhorn A, Sorkau E, Muller H, Rucknagel P, Nies DH, Fricke B: Camelysin is a novel surface metalloproteinase from Bacillus cereus. Infect Immun 2004;72:219-228.

Bacterial Peptidases

171

Gunther R, Stein A, Bordusa F: Investigations on the enzyme specificity of clostripain: A new efficient biocatalyst for the synthesis of peptide isosteres. J Org Chern 2000;65: 1672-1679. Hackbarth CJ, Chambers HF: blal and blaR I regulate beta-Iactamase and PBP 2a production in methicillinresistant Staphylococcus aureus. Antimicrob Agents Chemother 1993;37: I 144-1 149. Harrington OJ: Bacterial collagenases and collagen-degrading enzymes and their potential role in human disease. Infect Immun 1996;64: 1885-1891. Havarstein LS, Diep DB, Nes IF: A family of bacteriocin ABC transporters carry out proteolytic processing of their substrates concomitant with export. Mol Microbiol 1995; 16:229-240. Heath A, DiRita VJ, Barg NL, Engleberg NC: A two-component regulatory system, CsrR-CsrS, represses expression of three Streptococcus pyogenes virulence factors, hyaluronic acid capsule, streptolysin S, and pyrogenic exotoxin B. Infect Immun 1999;67:5298-5305. Heck LW, Morihara K, Abrahamson DR: Degradation of soluble laminin and depletion of tissue-associated basement membrane laminin by Pseudomonas aenlginosa elastase and alkaline protease. Infect Immun 1986;54:149-153. Hengge R, Bukau B: Proteolysis in prokaryotes: Protein quality control and regulatory principles. Mol Microbiol 2003;49: 1451-1462. Herwald H, Collin M, Muller-Esterl W, Bjorck L: Streptococcal cysteine proteinase releases kinins: A virulence mechanism. J Exp Med 1996;184:665-763. Herwald H, Morgelin M, Bjorck L: Contact activation by pathogenic bacteria: A virulence mechanism contributing to the pathophysiology of sepsis. Scand J Infect Dis 2003;35:604-607. Herwald H, Morgelin M, Olsen A, Rhen M, Dahlback B, Muller-Esterl W, Bjorck L: Activation of the contact-phase system on bacterial surfaces - A clue to serious complications in infectious diseases. Nat Med 1998;4:298-302. Hill HR, Bohnsack JF, Morris EZ, Augustine NH, Parker CJ, Cleary PI', Wu JT: Group B streptococci inhibit the chemotactic activity of the fifth component of complement. J Immunol 1988;141 :3551-3556. Hobden JA: Pseudomonas aeruginosa proteases and corneal virulence. DNA Cell BioI 2002;21: 391-396. Hofmann B, Schomburg 0, Hecht HJ: Crystal stnJchlre of a thiol proteinase from Staphylococcus oureus V-8 in the E-64 inhibitor complex. Acta Crystallogr A (Suppl) 1993;49: 102. Hosotaki K, Imamura T, Potempa J, Kitamura N, Travis J : Activation of protein C by arginine-specific cysteine proteinases (gingipains-R) from Porphyromonas gingivalis. Bioi Chern 1999;380:75-80. Hotson A, Chosed R, Shu H, Orth K, Mudgett MB: Xanthomonas type III effector XopD targets SUMOconjugated proteins in planta. Mol Microbiol 2003;50:377-389. Hu X, Zhu J, Srivathsan S, Pei 0: Peptidyl hydroxamic acids as methionine aminopeptidase inhibitors. Bioorg Med Chern Lett 2004;14:77-79. Hytonen J, Haataja S, Gerlach 0, Podbielski A, Finne J: The SpeB virulence factor of Streptococcus pyogenes, a multifunctional secreted and cell surface molecule with strepadhesin, laminin-binding and cysteine protease activity. Mol Microbiol 2001 ;39:512-519. Iiangovan U, Ton-That H, Iwahara J, Schneewind 0, Clubb RT: Structure of sortase, the trans peptidase that anchors proteins to the cell wall of StaphylococG1IS aureus. Proc Nat! Acad Sci USA 2001;98: 6056-6061. Imamura T, Banbula A, Pereira PJ, Travis J, Potempa J: Activation of human prothrombin by argininespecific cysteine proteinases (Gingipains R) from Porphyromonas gingivalis. J Bioi Chern 2001a; 276: 18984-18991. Imamura T, Pike RN, Potempa J, Travis J: Pathogenesis of periodontitis: A major arginine-specific cysteine proteinase from Porphyromonas gingivalis induces vascular permeability enhancement through activation of the kallikreinlkinin pathway. J Clin Invest 1994;94:361-367. Imamura T, Potempa J, Pike RN, Moore IN, Barton MH, Travis J: Effect of free and vesicle-bound cysteine proteinases of Porphyromonas gingivalis on plasma clot formation: Implications for bleeding tendency at periodontitis sites. lnfect lmmun I995a;63:4877-4882. Imamura T, Potempa J, Pike RN, Travis J: Dependence of vascular permeability enhancement on cysteine proteinases in vesicles of Porphyromonas gingivalis. Infect Immun 1995b;63: 1999-2003. Imamura T, Potempa J, Tanase S, Travis J: Activation of blood coagulation factor X by arginine-specific cysteine proteinases (gingipain-Rs) from Porphyromonas gingivalis. J Bioi Chern 1997;272: 16062-16067.

Potempa/Pike

172

Imamura T, Tanase S, Hamamoto T, Potempa J, Travis J: Activation of blood coagulation factor IX by gingipains R arginine-specific cysteine proteinases from POIphyromonas gingivalis. Biochem J 200 Ib;353:325-33 I. Imamura T, Travis J, Potempa J: The biphasic virulence activities of gingipains: Activation and inactivation of host proteins. Curr Protein Pept Sci 2003;4:443-450. Iriarte M, Cornelis GR: YopT, a new Yersinia Yop effector protein, affects the cytoskeleton of host cells. Mol Microbiol 1998;29:915-929. Isenberg HD: Pathogenicity and virulence: Another view. Clin Microbiol Rev 1988; I :40-53. Jagels MA, Travis J, Potempa J, Pike R, Hugli TE: Proteolytic inactivation of the leukocyte C5a receptor by proteinases derived from Porphyromonas gingivalis. Infect Immun 1996;64: 1984-1991. Janzon L, Lofdahl S, Arvidson S: Identification and nucleotide sequence of the delta-lysin gene, hlcl, adjacent to the accessory gene regulator (agr) of Staphylococcus aureus. Mol Gen Genet 1989; 2 I9:480-485. Johannsen DB, Johnston OM, Koymen HO, Cohen MS, Cannon JG: A Neisseria gonorrhoeae immunoglobulin A I protease mutant is infectious in the human challenge model of urethral infection. Infect Immun 1999;67:3009-3013. Juris SJ, Shao F, Dixon JE: Yersinia effectors target mammalian signalling pathways. Cell Microbiol 2002;4:20 I-2 I I. Kagawa TF, Cooney JC, Baker HM, McSweeney S, Liu M, Gubba S, Musser JM, Baker EN: Crystal structure of the zymogen form of the group A Streptococcus virulence factor SpeB: An integrinbinding cysteine protease. Proc Natl Acad Sci USA 2000;97:2235-2240. Kansal RG, McGeer A, Low DE, Norrby-Teglund A, Kotb M: Inverse relation between disease severity and expression of the streptococcal cysteine protease, SpeB, among clonal MIT I isolates recovered from invasive group A streptococcal infection cases. Infect Immun 2000;68: 6362-6369. Kansal RG, Nizet Y, Jeng A, Chuang WJ, Kotb M: Selective modulation of superantigen-induced responses by streptococcal cysteine protease. J Infect Dis 2003; 187:398-407. Kapur Y, Majesky MW, Li LL, Black RA, Musser JM: Cleavage of interleukin 1 beta (IL-I beta) precursor to produce active IL-I beta by a conserved extracellular cysteine protease from Streptococcus pyogenes. Proc Natl Acad Sci USA I993a;90:7676-7680. Kapur Y, Topouzis S, Majesky MW, Li LL, Hamrick MR, Hamill RJ, Patti JM, Musser JM: A conserved Streptococcus pyogenes extracellular cysteine protease cleaves human fibronectin and degrades vitronectin. Microb Pathog 1993b; 15:327-346. Kellner A, Robertson T: Myocardial necrosis produced in animals by means of crystalline streptococcal proteinase. J Exp Med 1954;99:495-503. Kessler E, Kennah HE, Brown SI: Pseudomonas pmtease. Purification, partial characterization, and its effect on collagen, proteoglycan, and rabbit corneas. Invest Ophthalmol Vis Sci 1977; 16:488-497. Kilian M, Reinholdt J, Lomholt H, Poulsen K, Frandsen EV: Biological significance of IgAI proteases in bacterial colonization and pathogenesis: Critical evaluation of experimental evidence. APMIS 1996; I04:321-338. Klein CD, Schiffmann R Folkers G, Piana S, Rothlisberger U: Protonation states of methionine aminopeptidase and their relevance for inhibitor binding and catalytic activity. J Bioi Chem 2003; 278:47862-47867. Komori Y, Nonogaki T, Nikai T: Hemorrhagic activity and muscle damaging effect of Pseudomonas aentginosa metalloproteinase (elastase). Toxicon 2001 ;39: I327- 1332. Koonin EY, Aravind L: Origin and evolution of eukaryotic apoptosis: The bacterial connection. Cell Death Differ 2002;9:394-404. Kosowska K, Reinholdt J, Rasmussen LK, Sabat A, Potempa J, Kilian M, Poulsen K: The Clostridium ramosum 19A proteinase represents a novel type ofmetalloendopeptidase. J BioI Chem 2002;277: I 1987-1 1994. Kramer RA, Dekker N, Egmond MR: Identification of active site serine and histidine residues in Escherichia coli outer membrane protease OmpT. FEBS Lett 2000;468:220-224. Kramer RA, Vandeputte-Rutten L, De Roon GJ, Gros P, Dekker N, Egmond MR: Identification of essential acidic residues of outer membrane protease OmpT supports a novel active site. FEBS Lett 200 I; 505:426-430.

Bacterial Peptidases

173

Kukkonen M, Lahteenmaki K, Suomalainen M, Kalkkinen N, Emody L, Lang H, Korhonen TK: Protein regions important for plasminogen activation and inactivation of alpha2-antiplasmin in the surface protease Pia of Yersinia pestis. Mol Microbiol 200 I;40: 1097-1 I I I. Kukkonen M, Suomalainen M, Kyllonen P, Lahteenmaki K, Lang H, Virkola R, Helander 1M, Holst 0, Korhonen TK: Lack of O-antigen is essential for plasminogen activation by Yersinia pestis and Salmonella enterica. Mol Microbiol 2004;51 :215-225. Kumagai Y, Konishi K, Gomi T, Yagishita H, Yajima A, Yoshikawa M: Enzymatic properties ofdipeptidyl aminopeptidase IV produced by the periodontal pathogen Porphyromonas gingivalis and its participation in virulence. Infect Immun 2000;68:716-724. Kumagai Y, Yajima A, Konishi K: Peptidase activity of dipeptidyl aminopeptidase IV produced by Porphyromonas gingivalis is important but not sufficient for virulence. Microbiol Immunol 2003; 47:735-743. Kuroda A, Sekiguchi J: Molecular cloning and sequencing of a major Bacillus subtilis autolysin gene. J Bacteriol 1991; 173:7304-7312. Labrou NE, Rigden DJ: The structure-function relationship in the clostripain family ofpeptidases. Eur J Biochem 2004;271 :983-992. Ladds G, Davey J: Identification of proteases with shared functions to the proprotein processing protease Krpl in the fission yeast Schizosaccharomyces pombe. Mol Microbiol 2000;38: 839-853. Lahteenmaki K, Kukkonen M, Korhonen TK: The Pia surface protease/adhesin of Yersinia pestis mediates bacterial invasion into human endothelial cells. FEBS Lett 2001;504:69-72. Langer T: AAA proteases: Cellular machines for degrading membrane proteins. Trends Biochem Sci 2000;25:247-251. Lathem WW, Grys TE, Witowski SE, Torres AG, Kaper 18, Tarr PI, Welch RA: StcE, a metalloprotease secreted by Escherichia coli 0157:H7, specifically cleaves CI esterase inhibitor. Mol Microbiol 2002;45:277-288. Lei B, DeLeo FR, Hoe NP, Graham MR, Mackie SM, Cole RL, Liu M, Hill HR, Low DE, Federle MJ, Scott JR, Musser JM: Evasion of human innate and acquired immunity by a bacterial homolog of CO II b that inhibits opsonophagocytosis. Nat Med 200 I;7: 1298-1305. Lei B, DeLeo FR, Reid SO, Voyich JM, Magoun L, Liu M, Braughton KR, Ricklefs S, Hoe NP, Cole RL, Leong JM, Musser JM: Opsonophagocytosis-inhibiting mac protein of group A Streptococcus: Identification and characteristics of two genetic complexes. Infect Immun 2002;70:6880-6890. Lei B, Liu M, Meyers EG, Manning HM, Nagiec MJ, Musser JM: Histidine and aspartic acid residues important for immunoglobulin G endopeptidase activity of the group A Streptococcus opsonophagocytosis-inhibiting Mac protein. Infect Imrnun 2003;71 :2881-2884. Li JY, Cui YM, Chen LL, Gu M, Li J, Nan FJ, Ye QZ: Mutations at the S I sites of methionine aminopeptidases from Escheri Sec-mediated translocation. Spa also contains a consensus carhoxyterminal sorting signal that consists of an LPXTG sequence motiffollowed by a hydrophobic stretch of 15-19 amino acids and a distal positively charged tail of 5-10 residues [29]. The hydrophobic/charged domain in the carboxy-terminus functions to retain Spa in the membrane after signal peptidase cleavage (P2

Bacterial lnvasins

185

NH 2

NH 2

セ セ L P X

2

-----.

T G

-GN-MN-

, ,

L-AJa

Gly5-L\YSJ , D-Ala

L P X

L

-----.

3

4

-----.

T G

G

P X T,

GN-MN-PP

L'Ala-----'

dセgャョ

,

L P X T,

-GN-MN

lセaャ。

,'

セMnhgiyUlゥケウ

セMnhgiyUlゥs

o

5

,

D-Ala

o

GN-MN-

, , ,

L-Ala

G1Y5- LjLys D-Ala

D-AJa

D-A1a

1M

B

+

j1 Mila

Fig. 2. The sorting reaction in S. aureus. PI precursor protein substrates that harbor an amino-tenninal signal peptide and a carboxy-terminal sorting signal are exported from the cytoplasm through the Sec translocase (1). The amino-tenninal secretion signal is cleaved by signal peptidase generating a P2 precursor, which is retained in the plasma membrane by the carboxy-terminal sorting signal (2). Sortase (SrtA) catalyzes a cleavage reaction between the threonine and glycine residues of the LPXTG motif, generating a thioester enzyme intermediate (3). The acyl-enzyme intermediate is resolved through nucleophilic attack by a free amine group on lipid II, resulting in amide linkage of the sortase substrate to the pentaglycine cross-bridge (4). The mature surface protein is incorporated into the cell wall through a transglycosylation reaction (5). 1M = Inner membrane.

precursor). Sortase catalyzes a transpeptidation reaction between the threonine and glycine residues of the LPXTG motif, where a proteolytic cleavage event links the threunyl carbuxyl group tu an active site cysteine, generating an acylenzyme intermediate through thioester linkage [30]. The carboxyl group ofthreonine is then amide linked to the amino group of the pentaglycine cross-bridge in the murein tetrapeptide segment of the lipid II cell wall precursor [31]. The reaction product is incorporated into new peptidoglycan polymers through transglycosylation, resulting in the mature cell wall-anchored Spa polypeptide. The general staphylococcal sorting reaction is represented in figure 2. The identification of cell wall-anchored surface proteins based on carboxyterminal sequence analysis has revealed the potential for numerous virulenceassociated factors, including C5 peptidase in Streptococcus pyogenes, internalin A

Cambronne/Schneewind

186

in 1. monocytogenes, and neuraminidase in Streptococcus pneumoniae [28]. Variations on the consensus sorting reaction have also emerged, which include multiple sortase enzymes encoded by the same organism that recognize alternate sorting signals, such as the recognition of the NPQTN consensus sorting signal by SrtB in S. aureus [32]. Further, evidence suggests that the expression of sortase enzymes is environmentally regulated, promoting the display of a particular set of surface proteins under specific conditions. 1. monocytogenes is a food- and water-borne pathogen that causes infections ranging from gastroenteritis to septicemia. It is an intracellular pathogen that employs a particularly interesting mechanism for migration through host tissue. The bacterium requires at least two surface factors for entry into cultured cells. Internalin A (InlA) and internal in B (InlB) each contain Sec-mediated amino-terminal signal sequences but are recruited to the bacterial surface by two different mechanisms [33]. 1. monocytogenes harbors two sortase genes, srtA and srtB. SrtA is required for the cell wall anchoring ofInlA, which contains a consensus LPXTG sorting signal. Deletion of the srtA locus results in a defect in invasiveness similar to that of an inlA mutant [34]. InIB contains carboxy-terminal repeat regions that promote a noncovalent interaction with lipoteichoic acids in the cell wall peptidoglycan [35]. InlA binds to the E-cadherin receptor on epithelial cells, while InlB interacts with the complement receptor gC 1qR, glycosaminoglycans, and the tyrosine kinase receptor Met [36-38]. Activation of signal transduction cascades promotes phagocytosis and the bacterium is enveloped in a phagocytic vesicle. To combat the acidification of the phagocytic vacuole, the bacterium expresses the enzymes listeriolysin 0 (LLO) and phosphatidylinositol phospholipase C (PleA), which are secreted and promote degradation of the vacuolar membrane [39,40]. This event allows for escape from the phagocytic vacuole and promotes bacterial multiplication in the host cell cytoplasm. Listeria expresses another factor ActA, a membrane protein exposed on the bacterial surface. ActA recruits the host Arp 2/3 complex resulting in nucleation of actin filaments at the surface of the bacterium [41, 42]. ActA acts as a molecular mimic, functioning in a similar fashion to the WASP family of proteins. The WASP proteins are activated through binding of cellular GTPases, and conformational changes promote Arp 2/3 complex recruitment [43]. The assembly of an actin tail propels the bacterium through the cytoplasm, generating pseudopod-like extensions that promote phagocytosis by neighboring cells, resulting in the formation of a double membrane vacuole in the neighbor cell. In addition to the secretion of LLO and PIcA, phosphatidyleholine phospholipase C (PlcB) has been implemented in the escape of the bacterium from this specialized vacuole [44]. The intracellular growth cycle of the bacterium has been shown to result in localized tissue destruction with minimal exposure to components of the immune system.

Bacterial lnvasins

187

Invasive Strategies of Gram-Negative Pathogens

Gram-negative pathogens have devised an array of mechanisms to promote colonization. Most of these strategies incorporate the modification of a generalized secretion pathway to either promote the display of a surface molecule for colonization, or deliver effector molecules beyond the bacterial envelope. Specialized secretion systems may be generally divided into two categories, those that promote the release of diffusible protein factors to the surrounding environment and systems that promote the delivery of effector proteins directly into the cytosol of target cells. There are currently five specialized secretion systems described for gram-negative bacteria (type I-V) that appear dedicated to virulence. Type I secretion incorporates a Sec-independent process to deliver toxin to the extracellular space without a periplasmic intermediate. Type II secretion, the main terminal branch of the general secretory pathway (GSP), represents a two-step translocation mechanism where factors secreted by the Sec pathway are transported by a protein complex that contains a characteristic outer membrane secretin. The type III secretion mechanism is a Secindependent translocation process that involves the direct delivery of effector molecules from the bacterial cytoplasm to the cytosol of a target cell through a specialized channel or needle complex. Type IV secretion systems are similar to bacterial conjugational systems and harbor the ability to transfer proteins ancl/or nucleic acids into a target cell using either a one- or two-step translocation process. Type V secretion represents an alternate terminal branch of the GSP. Often referred to as the autotransporter mechanism, type V substrates are secreted by the Sec pathway and contain information in their carboxy-termini that promotes incorporation in the outer membrane and delivery of the aminotermjnal domain outside of the cell. Each of these systems are discussed in detail below and figure 3 is a representation ofthe basic features associated with these secretion mechanisms. Surface Proteins The expression of a molecule on the smface of the bacterium, not unlike the display of surface proteins in gram-positive pathogens, represents a mechanism for colonization in some gram-negative bacteria. A prototypical example of this mechanism is the display of the factor invasin in Yersinia pseudotuberculosis and Yersinia enterocolitica. The expression of invA in a noninvasive strain of E. coli results in the phagocytosis of the bacterium by cultured mammalian cells [45]. Invasin is a modular protein that harbors an amino-terminal outer membrane localization domain, as well as an extracellular carboxy-terminal domain that consists of repeats of an IgG-like fold, and an adhesive tip [46, 47]. It has been determined that invasin binds to 13, integrin receptors localized on

Cambronne/Schneewind

188

OM

cw 1M

Fig. 3. Basic features of the five classes of gram-negative protein secretion systems. (I) The type 1 mechanism involves a single, Sec-independent translocation event that incorporates an inner membrane (1M) ABC transporter for energy generation. (II) The type II mechanism represents a two-step translocation process, where signal-bearing precursor proteins are transported to the periplasm via the Sec pathway. Mature substrates are transported through an outer membrane (OM) secretin. (III) The type III mechanism incorporates a single translocation step to transport substrates from the bacterial cytoplasm into the cytosol of eukaryotic cells. The substrate is transported through a basal body complex, an outer membrane secretin, and a needle complex that penetrates the target cell membrane. (Iv, left) The type IV secretion system is employed to transfer substrates into host cells. This process requires the assembly of a pilus structure at the outer membrane, a core assembly in the peri plasm, and inner membraneassociated ATPases. (Iv, right) Pathogens may also employ the type IV mechanism for secretion of diffusible toxins to the environment in a Sec-dependent manner. (V) Autotransporters are secreted by the type V pathway. A typical substrate is translocated to the periplasm by the Sec translocase. Insertion into the outer membrane promotes the secretion of the amino-terminal passenger domain. Autoproteolysis releases the diffusible mature protein. Localization of energy generating enzymes are indicated by *. CW = Cell wall peptidoglycan.

the apical surface of M cells, located amongst the follicle-associated epithelia and lymphoid follicles of the small intestine, commonly referred to as Peyer's patches [48,49]. M cells sample contents of the intestinal lumen and transport particles contained in vesicles to the basolateral surface, which is rich in immune cells such as macrophages and polymorphonuclear leukocytes. The binding of invasin to M cells may therefore represent an early mechanism involved in the Yersinia infection process, as the bacteria have a specific tropism for lymphoid tissues. Numerous examples of adhesion factors have been identified, often associated with the protein subunits localized in the tip of pili or fimbriae.

Bacterial lnvasins

189

Examples include the PapG adhesin of type I pili in E. coli, and the major pilin subunit PilA on the type IV pili of P aeruginosa [50, 5]].

Type I Secretion The type I secretion mechanism involves the one-step translocation of a secretion substrate in a Sec-independent manner [52]. This mechanism is employed for the secretion of diffusible toxins into the extracellular space. The type I mechanism has been demonstrated for the secretion ofa-hemolysin (HlyA) in E. coli, as well as for the secretion of Bordetella pertussis adenylate cyclase and P aenlginosa protease [53]. In each instance, the toxin is recruited to a translocation complex that assembles upon association with the substrate. The type I secretion system is relatively simple in architecture, consisting ofonly three factors, each required for transport of the substrate. A characteristic feature of the system is the presence of an ATP-binding cassette (ABC) protein transporter [52, 54]. ABC transporters are inner membrane proteins that are found in a wide range of organisms, including gram-positive bacteria, lower eukaryotic, and mammalian cells, and are normally associated with the transport of small molecules. The secretion of HlyA has been extensively studied and the synthesis of the prohemolysin precursor protein (proA) requires a lipid modification for activation to mature HlyA [55]. The cytosolic factor HlyC as well as an acyl-carrier protein (ACP) are required for the myristoylation or palmitoylation of two lysine residues [56]. HlyC acts as an acyl-transferase for this process. Although this lipid modification step is required for the hemolytic activity of HlyA, this event is not required for the type I-dependent secretion of HlyA [57]. After modification, HlyA binds to the ABC transporter HlyB at the inner membrane [58]. The sequence information required for secretion of HlyA is contained in the polypeptides carboxy-terminal 48 amino acids, and unlike signal sequences in Secmediated substrates, the signal sequence of type I substrates is not cleaved after translocation [59, 60]. The HlyB transporter associates with a second factor HlyD independent of substrate binding. The HlyD protein spans both the inner and outer membrane and trimerization of HlyD in the presence of HlyB bound to HlyA resulls in the recruitment of the outer membrane protein TolC [58]. Each subunit of the trimeric TolC contains an amino-terminal l3-sheet domain that inserts into the outer membrane. A second carboxy-terminal a-helical domain extends deep into the periplasmic space and forms a barrier between the periplasm and the amino-terminal pore-like structure [61]. The binding ofATP to HlyB in the presence of HlyA may result in the specific recruitment ofTolC by HJyD. The HlyBfHlyDffolC complex then supports the HJyB mediated translocation of HlyA through successive rounds of ATP hydrolysis [62], resulting in the delivery of HlyA to the extracellular space. Eleven tandem glycine-rich repeats (LXGGXGND) contained in the carboxy-terminus ofHlyA are required

Cambronne/Schneewind

190

for calcium binding. Calcium-bound HlyA is competent for insertion in the host cell membrane, and results in pore-mediated leakage of the target cell [63]. Type II Secretion The GSP represents the primary route for translocation of polypeptides to the extracellular space among gram-negative bacteria. The type II secretion mechanism represents the archetype for protein secretion in the GSP, and has been designated the main terminal branch. The type II pathway is associated with the secretion of virulence factors in several bacterial pathogens. Alternate GSP branches include the secretion of autotransporters (type V secretion), the chaperone/usher-mediated assembly of P or type I pili in E. coli, the assembly of type IV pili in P aeruginosa and Neisseria gonorrhea, and the assembly of curli in E. coli [64]. The factors required for extrusion of filamentous bacteriophage from the bacterial envelope also share conserved components with the GSP [65]. One common feature associated with all of these strategies is the requirement for the Sec-mediated translocation of secretion substrates to the periplasm. The type II secretion pathway therefore represents a two-step translocation process, incorporating distinct secretion reactions for translocation across the inner and outer membranes. The type II-dependent secretion ofpullulanase in Klebsiella oxytoca is a wellstudied example of this secretion mechanism. Pullulanase (PulA) is a lipoprotein of the a.-amylase family that enzymatically degrades the complex carbohydrate pullulan to maltotriose subunits, a substrate that may be transported into the bacterium [66]. The secretion of PulA requires the products of at least 25 genes, 14 of which are specifically involved in the translocation of PulA beyond the outer membrane [67]. After secretion to the periplasm through the Sec pathway, The PulA precursor is subjected to diacyl glyceride modification and cleaved by signal peptidase [68]. The lipid-modified PulA is retained in the outer leaflet of the inner membrane by an aspartyl residue located at the amino-terminus of the mature polypeptide. Factors required for the type II-dependent translocation step are localized within several compartments. A cytoplasmic ATPase (GspE) associates with the inner membrane through interaction with a second factor (GspL), an inner membrane protein that harbors a carboxy-terminal cytoplasmic domain [69, 70]. This interaction, coupled with ATP hydrolysis by GspE, may provide the energy required for PulA transport. Four additional integral membrane proteins GspC, GspF, GspM, and GspN are thought to assemble into a basal body complex, since the factors harbor carboxy-terminal domains that extend into the periplasm [71]. A characteristic feature of the type II apparatus is the requirement for periplasmic pseudopilin proteins. These factors, all harboring prepilin signal sequences, are secreted by the Sec pathway and processed by the inner membrane-associated prepilin peptidase GspO, which will also N-methlyate

Bacterial lnvasins

19J

the pseudopilin subunits [72]. Five pseudopilin factors, GspG, GspH, GspI, GspJ, and GspK, are processed in this manner, and assemble into a channel-like pilus structure linking components of the inner and outer membranes [73]. GspD is an outer membrane secretin required for the export of type II substrates. GspD is inserted into the outer membrane and assembles into a dodecameric channelforming structure, a process that requires the outer membrane chaperone GspS [74]. Functional homologs of the GspD secretin are conserved amongst most of the alternate branches of the GSP and the GspD secretin is also conserved among the type III secretion systems (see below). The type II-mediated export of PulA may occur through its association with the basal body complex, with signal recognition most likely residing in the secondary or tertiary structure of the secretion substrate after folding in the periplasm. PulA is transported to the outer membrane secretin GspD and exported to the extracellular space. Similar type II secretion mechanisms have been identified for the release of exotoxin A, elastase, and phospholipase C in P aentginosa [71]. The secretion ofAB-type holotoxins is also mediated by a type II-dependent process. This class of toxins includes the cholera toxin of Vibrio cholerae, E. coli enterotoxin, and the Shiga-like toxins of E. coli and Shigella dysenteriae [75]. Cholera toxin is composed of two separate polypeptides CtxA and CtxB. Secdependent secretion ofthe subunits to the periplasm results in proteolytic cleavage of signal peptides and the formation of an intramolecular disulfide bond in CtxA prior to cleavage. The CtxB subunits assemble into a pentameric ring structure and bind the carboxy-terminal domain of the CtxA subunit, generating the CtxA1-CtxB s holotoxin [76]. Secretion of the holotoxin requires components of the eps gene cluster, which encodes several factors homologous to the type II secretion system in Klebsiella [77]. Export of the holotoxin will result in the binding of the CtxBs subunits to a G M1 ganglioside on the surface of intestinal epithelial cells [78]. Reduction of the disulfide in CtxA by host cytosolic thioredoxin promotes the release of the mature CtxA toxin from the CtxB s pentameric ring, where it will function to activate host cell adenlyate cyclase, resulting in the massive cellular fluid loss associated with the diarrhea in cholera disease [79].

TYpe III Secretion The delivery of polypeptides from the bacterial cytoplasm directly into the cytosol of target host cells without the generation of an extracellular intermediate is the hallmark feature of the type III secretion system [80]. Effector proteins that are translocated into host cells harbor enzymatic activities that manipulate cellular processes of the eukaryotic host, resulting in a variety of processes that culminate in perpetuation of the bacterium at the infection site [81]. The type III secretion mechanism was first characterized in pathogenic

Cambronne/Schneewind

192

Yersinia species, but has subsequently been identified and extensively studied in various pathogens including enteropathogenic E. coli (EPEC), P aeruginosa, Salmonella enterica and Shigella flexneri [82]. Analysis of genetic information has revealed the potential for type III systems in several other gram-negative bacteria, and thus may represent a highly conserved pathogenic strategy [83]. Even though the process of injection of virulence factors by the type III pathway is a recently described phenomenon, the type III secretion apparatus appears both structurally and functionally similar to the basal body of the flagellar secretion system in gram-negative bacteria. In fact, the flagellar secretion system is now considered a type III pathway and recent observations suggest that the flagellar export system may also support the secretion of virulence factors [84]. Type III secretion systems most likely evolved from the flagellar machinery to support colonization in new nutrient-rich environments such as those found in higher eukaryotes. Yersinia species employ the type III pathway to maintain an extracellular lifestyle in the lymphoid tissues of their mammalian hosts and cause a variety of diseases ranging from bubonic plague in Yersinia pestis to acute enteritis in Y enterocolitica. This is accomplished through the type III injection of effector proteins called Yops (Yersinia outer proteins) into host macrophages, resulting in prevention of phagocytosis and eventual apoptotic death of the host cell [85]. With the exception of the assembly of the secretion apparahls, the translocation of type III secretion substrates represents a Sec-independent process. The type III secretion system consists of three principle components, an inner membraneassociated basal body, an outer membrane secretin, and an extracellular needle complex. Assembly of a functional Yersinia type III apparatus requires the products of at least 21 ysc (Yersinia secretion) genes [86-88]. Eleven of these genes are conserved amongst other type III systems, including nine that are conserved with the flagellar basal body [89]. In general, the Yersinia type III secretion apparatus must be assembled in a similar fashion to the flagellar secretion system, where assembly of the basal body complex precedes any substrate delivery. The initiation of the assembly of the basal body complex in Yersinia likely begins with membrane insetiion of the FliF homolog YscJ, after Sec-mediated translocation [86, 90, 91]. This event will allow the association of inner membrane proteins YseD, YscR, YscU, YscV and accessory factors to form the basal body complex. YscN is homologous to the FliI ATPase in flagellar secretion and contains the Walker boxes A and B, which are characteristic conserved ATP-binding domains [92]. YseN is predicted to provide the energy for the transport of type III secretion substrates and is required for Yop secretion. YseC is homologous to the GspD secretin involved in the transport of molecules in the type II pathway, and requires the outer membrane lipoprotein YscW for its localization and for the formation of the characteristic dodecameric rings in the outer membrane, resulting in outer membrane

Bacterial lnvasins

193

channel formation [93, 94]. Accessory factor association between the basal body complex and the YscC secretin provides a conduit between the two components of type III secretion system, an assembly step that is not conserved with flagellar secretion. Secretion ofthe factors YscF, YscO, YscP, and YscX promote the assembly of the needle complex [95-97]. Although the type III needle complex remains to be isolated from Yersinia, needle complexes have been pmified from Salmonella, Shigella, and E. coli [98-100]. The YscF protein has been determined to be the main component of the needle complex, where the protein multimerizes in a right-handed helical fashion. The YscF homolog MxiH of Shigella displays 5.6 subunits per tum, and is polymerized from the distal tip into a conduit as long as 50 nm with a width of 7 nm and a central tube of 2-3 nm [IO I, 102]. YscO and YscP are secreted by the basal complex. YscP has recently been suggested to participate in substrate recognition, as yscP mutant strains secrete an increased amount of the needle component YscF but fail to secrete Yop proteins in vitro. Mutations in the amino-terminus ofYscU suppress the yscP mutant phenotype, reducing the amount of secreted YscF to wild-type levels and restoring the secretion of Yop proteins [95]. These results suggest that YscO, Yscp' and YscU may control type III secretion at the level of substrate specificity, allowing for a switch between the secretion of structural components and the delivery of Yop substrates, similar to the switch between hook and filament proteins in the flagellar apparatus [89]. Assembly of the YscF needle complex would then represent the final step of assembly and provide a switch for the recognition of type III secretion substrates and the delivery of effector Yop proteins. The hydrophobic natme of the YscF polymer has been predicted to provide a mechanism for the piercing of the host cell cytoplasm [96]. An alternate hypothesis suggests that three secretion substrates, YopB, YopD, and LcrY, each required for the translocation ofYop proteins, form a translocation pore in the host cell membrane allowing for subsequent delivery of effector proteins [80, 103-106]. Y enterocolitica secretes 14 polypeptides via the type III pathway. One cmious feature of each of these proteins is that they do not contain any amino acid sequences that would suggest the presence of a conserved type III secretion signal. Experiments performed using reporter proteins have revealed the presence of minimal secretion information contained in the amino-terminal 8-15 residues [107-110]. The nature of the minimal secretion information remains controversial. Scanning mutagenesis studies employed to determine the residues required for secretion of YopE revealed that no specific residues were necessary. Further, introduction of frameshift mutations in the minimal signal did not affect the secretion of reporter fusion constructs [107]. These results prompted the hypothesis that the minimal signal information is actually

Cambronne/Schneewind

194

contained in the mRNA rather than the protein. The yop mRNA might therefore recruit a translational complex to the type III apparatus promoting a cotranslationa I secretion mechanism. The nature of the 5' mRNNamino-terminal signal hypothesis has been highly contested however, and independent studies suggested that a mutant that generated multiple mutations in the mRNA sequence without affecting the amino acid sequence of the protein was indeed secreted [109]. This implied that the amino acid sequence, rather than the mRNA, contained the information required for type 1II secretion. Construction of a synthetic amphipathic amino-terminal signal that contained alternating serine and isoleucine residues between positions 2 and 9 ofYopE also supported secretion [109]. Recent observations in the minimum secretion signals of YopE and YopQ have again resurrected the mRNA secretion signal argument, as it was discovered that minimum signals, such as the 1-10 positions ofYopQ, do not tolerate frameshifts unless a downstream suppressor region ofmRNA is included that contains codons 11-13 [110]. Further, single substitutions in codons 2 and 10 caused a defect in the secretion reporter fusions in the context of 10 but not 15 codons. Finally, multiple mutations in the wobble positions of yopQ 1-10 did not support the secretion of the reporter, again suggesting that mRNA rather than protein sequences initiate the transport of substrates via the type III pathway [110]. Beyond the context of the amino-terminal minimal secretion signal, experimental evidence suggests that type III substrates may require a second signal for their injection into the cytosol of host cells, and the presence ofSyc (specific Yop chaperone) proteins may be required for the injection process [105, 111]. In general, Syc proteins are small acidic proteins that form dimers in the bacterial cytoplasm. Each chaperone appears to specifically bind a partner effector Yop protein in the cytoplasm and structures have been determined for secretion substrate/ chaperone complexes [112]. Studies that examined the role of both the ammoterminal and chaperone-mediated secretion signals demonstrated that a defective secretion signal, when linked in context to the full length YopE protein, was secreted in an SycE-dependent manner, suggesting that the chaperone mediated secretion signal does not require the presence of a functional amino-terminal signal [113]. This prompted the hypothesis that Yop proteins harbor two independent secretion signals, the first required for initiation of the substrate into the type III pathway, and the second for injection into host cells. Y. enterocolitica transports a class of at least six factors into the cytosol of the host cell, each of which harbors an enzymatic function. All of these factors, which include YopE, YopH, YopM, YopO, YopP, and YopT, share sequence homology to proteins of eukaryotic origin, suggesting the pathogen evolved these strategies through intimate interaction with the host over time. Although all of the type III pathogens share a conserved mechanism for the delivery of these

Bacterial lnvasins

195

effector proteins, the effector proteins themselves mayor may not be conserved between pathogens. YopE is only cytotoxic to HeLa cells when injected via the type III pathway. It is a characteristic GTPase-activating protein (GAP) that acts upon the Rho family of eukaryotic GTPases. YopE inactivates RhoA, Rac I, and CDC42 by accelerating the conversion of GTP to GDP in these factors [I14, lIS]. This mechanism results in an inhibition of actin polymerization at the site of bacterial contact. YopH is a protein tyrosine phosphatase (PTPase) that is involved in the dephosphorylation offocal adhesions [116]. The amino-terminal domain of YopH appears to be a targeting domain that binds to p130cas and focal adhesion kinase (FAK) [117]. The carboxy-terminal domain harbors the phosphatase domain and acts specifically to dephosphorylate these substrates, resulting in an interruption of stress fiber formation [118]. YopM is required for virulence in mice and has been suggested to target to the nucleus and may influence transcription in the host cell, inhibiting inflammatory cytokine production [119, 120]. YopO is similar in sequence to RhoA kinase. YopO functions as an autophosphorylating serine threonine kinase that is activated in vitro through binding to actin [121]. The protein is believed to phosphorylate the Rho family of GTPases and enhance the inhibition of actin polymerization [122]. YopP acts as an inhibitor oflKB in the NF-KB pathway and also inhibits the MAP kinase pathway [123-125]. It has been reported that YopP is a cysteine protease that may function through a protein degradation pathway [126]. The cumulative effects of disrupting the NF-KB and MAP kinase pathways result in inhibition of the pro inflammatory response, thus preventing the production of the cytokines TNF-a and IL-8 [123, 125]. YopP also induces apoptosis in macrophages, which is likely to be a cumulative result of the failure to activate the NF-KB signaling pathway and through the cleavage of Bid, a proapoptotic member of the Bcl-2 family [127]. YopT is also a cysteine protease that has been demonstrated to cleave RhoA, Racl, and CDC42 at their carboxy-termini, sites that are prenylated for membrane anchoring [128]. Cleavage releases the factors from the membrane resulting in a defect in actin polymerization at the site of bacterial contact. EPEC, a food- and water-borne pathogen that is a causative agent of human infantile diarrhea, uses the type III pathway to establish attaching and effacing lesions on intestinal epithelium [129, 130]. These bacteria inject a protein, translocated intimin receptor (Tir), which is subsequently displayed on the surface of the gastric epithelial cell [131]. The bacterial cell displays a ligand for this receptor on its outer membrane, called intimin. Interaction between the two factors results in tight binding between the bacterium and host cell. This event coupled with the cumulative effects of the type III injection of other factors will promote actin pedestal formation at the site of contact, allowing extracellular colonization and destruction of surrounding tissue.

Cambronne/Schneewind

196

S. enterica serovar spp. cause a variety of diseases in humans and animals, ranging from acute food poisoning and gastrointestinal inflammation to typhoid fever and septicemia. In general pathogenic Salmonella species are foodand water-borne pathogens that have a tropism for the intestinal epithelium. S. enterica serovar spp. harbor the genes encoding two separate type III secretion systems on their chromosome. The first system, designated Salmonella pathogenicity island 1 (SPI-l), is employed to invade nonphagocytic epithelial cells [132]. Salmonella uses the SPI-I type III pathway to inject several effector molecules that leads to a massive reorganization of actin fi laments promoting the formation of membrane rumes and eventual phagocytosis. Similar to effector proteins in Yersinia, many Salmonella effectors target the signaling processes governing actin polymerization. SipA stabilizes F-actin through binding to the T-plastin protein [133]. SopE and SopE2 function as guanine nucleotide exchange factors (GEF) to activate Rac-l and CDC42 [134,135]. In order to promote recovery of the cytoskeleton, SptP is injected. SptP is a multifunctional enzyme that contains an amino-terminal YopE-like GAP domain and a carboxy-terminal YopH-like tyrosine phosphatase domain. SptP counteracts the enzymatic effects of SopE and SopE2 by downregulating Rac-l and CDC42 [136, 137]. The second type III system located at SPI-2 appears to manipulate vesicular trafficking, allowing for perpetuation of the bacterium in a specialized vacuole, and is required for systemic infections [138]. Pathogenic Shigella species are typically water-borne pathogens that are a causative agent ofbacillary dysentery, an infection of the colon. Shigella species utilize a type III secretion for the invasion of epithelial cells. Shigella are believed to enter epithelial cells from the basolateral surface. After engulfment by intestinal M cells and presentation to lymphoid macrophages, Shigella secretes an apoptotic factor IpaB which allows the bacterium to spread to adjacent cells [139]. Shigella also employs the type III pathway to promote phagocytosis by the epithelial cell through the cumulative effects ofIpaA, IpaB, IpaC, and IpaD [140]. Unlike Salmonella, Shigella escape from acidified vesicles and reside in the host cell cytoplasm. An outer membrane protein IcsA nucleates actin polymerization through the binding to N- WASP and the fOll11ation of the Arp213 complex [141, 142]. Production of cytoskeletal filaments at the pole of the bacterium propels the organism into neighboring cells, similar to the process described for L. monocytogenes. TYpe IV Secretion The type IV secretion mechanism is employed for a wide range of functions in gram-negative bacteria. Several species utilize the type IV mechanism for interbacterial conjugative transfer of mobilized genetic elements. Pathogenic Agrobacterium tumefaciens employs the type IV system for the transfer of

Bacterial lnvasins

197

tumorigenic DNA and protein into host plant cells, and several vertebrate pathogens such as Bnlcella spp., B. pertussis, Helicobacter pylori, and Legionella pneumophila use a modified type IV secretion system for the secretion of toxins or the delivery of effector proteins into the host cell [143, 144]. There is evidence to suggest that the secretion of substrates through the type IV apparatus requires a Sec-dependent translocation step, such as in the secretion of pertussis toxin; however, specialized systems such as those for H. pylori and L. pneumophila may bypass this requirement [145]. All type IV systems represent a modification of the conjugative transfer system found in strains of E. coli. In general, the mechanism involves the assembly of a secretion apparatus with a pilus-like projection that will provide intimate contact between the donor and recipient cell [146]. The transfer of DNA from bacterium to host in the pathogen A. tumefaciens represents the archetype for the type IV secretion pathway. Substrate translocation requires the products of the VirB-encoded system: VirB 1-11 and VirD4 [147]. The mechanisms of VirB-mediated type IV transport have been extensively studied, and the system is currently employed as the general model for the type IV mechanism in animal pathogens [144]. Evidence suggests that the type IV apparatus is assembled to extract the major pilin subunit VirB2, a cyclic polypeptide, through the outer membrane [148]. The secretion and processing of the pilin subunits is a Sec-dependent process. VirB2 forms a pilus through multimerization and contains a second minor pilin subunit VirB5. Pilin subunits interact with an outer membrane lipoprotein VirB7 and are thought to assemble at the outer membrane [149]. VirB6 is an inner membrane protein that may provide the connection between components of the inner and outer membrane, as well as guide assembly of the periplasmic core [150]. VirB7 also interacts with VirB9, an outer membrane component, and VirB8, a muramidase localized in the periplasm that may provide for organization of the complex through wall peptidoglycan [149, 151]. Energy for the transport of type IV substrates is provided by the activities of three separate ATPases. The VirB4 dimer is localized in the inner membrane. A second inner membrane ATPase, hexameric VirB 11 assembles into a ring structure and may provide a route for translocation of type IV substrates [152, 153]. VirB 11 also interacts with the periplasmic core component VirB 10. The third ATPase VirD4, also called 'coupling protein', is localized in the bacterial cytoplasm and is involved in substrate recognition [144]. The ptl system in B. pertussis represents an interesting link between the type II and type IV secretion pathways. B. pertussis is the causative agent of whooping cough, where pertussis toxin, an AB-type holotoxin, is the primary virulence determinant. Pertussis toxin is exported by the Ptl type IV secretion apparatus. The Ptl system appears functionally distinct from other type IV

Cambronne/Schneewind

198

secretion systems. Rather than supporting injection into host cells, the Ptl system exports pertussis toxin to the extracellular space, where the toxin is available to associate with the target cell membrane [145]. The mature enzyme acts as an ADP-ribosylating factor of G proteins in the host cell. Unlike the T-DNA translocation process, the PtxA and PtxB subunits are translocated to the periplasm, where they are processed and assembled into the PtxAt-PtxB s holotoxin [154]. This protein complex then becomes a substrate for type IV-mediated export. Nine structural components of the Ptl system are homologous to the VirB system, where PtxA represents the major pilin subunit. The system also contains two membrane-associated ATPases, PtiC and PtJH, which are homologous to VirB4 and VirB 11, respectively. Recent discoveries have provided evidence that type IV secretion systems are competent in the delivery of effector proteins directly into the cytosol of host cells. H. pylori is a causative agent of several gastrointestinal syndromes ranging from peptic ulcers to MALT lymphoma and adenocarcinoma. Pathogenic strains of H. pylori harbor the CAG pathogenicity island which encodes a VirB-like type IV secretion system [155]. CagA is translocated by the type IV pathway into the host cell cytosol where it becomes tyrosinephosphorylated and proteolytically processed to a carboxy-terminal phosphorylated fragment. The injection of CagA results in a change in the phosphorylation state of associated host cell factors, and is required for virulence [156]. L. pneumophila and Brucella species require type IV secretion systems for survival in intracellular vacuoles [157]. L. pneumophila is the causative agent of Legionnaire's disease, a severe respiratory pneumonia. L. pneumophila targets alveolar macrophages where it employs the Dot/Icm type IV secretion system for intracellular survival. The Dot/lcm transporter is more distantly related to the VirB system, but is homologous to the IncI conjugation system in S. jlexneri, and is competent for conjugational transfer of DNA [158]. L. pneumophila bypasses destruction mediated by the endocytic pathway by creating an endoplasmic reticulum-like vacuole, presumably through the injection of effector molecules into the host cytosol [157]. RalF was the first factor identified to be an effector substrate. RalF is a guanine nucleotide exchange factor that functions to activate the ADP ribosylation factor (ARF) family of GTPases [159]. The factor has been localized on the surface of the Legionella-containing vacuole in a Dot/lcm-dependent manner, and is required for the early recruitment of ARF 1, but is not required for intracellular survival of the bacterium. A second factor LidA has recently been identified to be exported in a DotlIcm-dependent manner and localizes to the phagosomal surface [160]. It has been hypothesized that this factor may function as a gatekeeper for the premature release of other factors. It is not yet clear how L. pneumophila modulates the type IV pathway to support the export

Bacterial lnvasins

199

of nucleic acid/protein hybrids or effector protein substrates under different environmental stimuli. The requirement for at least 24 genetic loci to promote intracellular survival suggests that the Dot/Icm transporter may be far more sophisticated than the VirB-like type IV systems [157].

Type V Secretion (Autotransporters) The autotransporter pathway represents an alternate branch ofthe GSP that serves as a simplified mechanism for the translocation of substrates out of the cell. Analysis of various genomes suggests that the autotransporter mechanism is widely conserved across gram-negative bacteria [83]. Rather than a requirement for factors in the periplasm or outer membrane translocases, the autotransporter secretion substrate harbors infonnation in its carboxy-terminus for insertion into the outer membrane, and for translocation of the amino-terminal domain out ofthe bacterium. Pathogenic species of Neisseria secrete Igal protease using the type V mechanism to promote survival in interstitial fluids. The activated enzyme, once exported will function to degrade secretory antibodies [161]. Iga I protease is synthesized as a preproenzyme that harbors an aminoterminal signal sequence to initiate its translocation across the inner membrane through the Sec pathway. The signal sequence is cleaved by signal peptidase, and the carboxy-terminal domain folds into a r3-barrel structure promoting insertion of the proenzyme in the outer membrane [162]. Insertion in the outer membrane generates a porin-like channel for the export of the amino-terminal passenger domain. Transport of the amino-terminal domain through the channel promotes an autoproteolysis event, cleaving the proenzyme between the N- and C-tenninal domains at a proline residue [163]. This proteolytic event will result in the release of a diffusible active enzyme. Examples of autotransport have also been described for the Hap adhesin of Haemophilus injluenzae, the IcsA protein of Sjlexneri, and the adhesin YadA of Y enterocolitica. The IcsA autotransporter represents a modification of the type V pathway, where an outer membrane serine protease SopA is required for the cleavage of the IcsA passenger domain, which promotes proper actin cytoskeletal nucleation [164]. YadA in Y enterocolilica may represent another modification of the type V pathway, where the amino-terminal passenger domain is delivered to the extracellular space, but is not cleaved from the carboxy-tenninal r3-barrel [165]. Further, YadA assembles into trimers in the outer membrane and the amino-terminal heads assume a lollipop-like structure extending from the narrow stalk domain. YadA is involved in the resistance to complement mediated lysis. A second variation of the type V pathway involves a two-component secretory system, where the synthesis and secretion of an enzyme substrate requires a single outer membrane transporter for delivery to the extracellular

Cambronne/Schneewind

200

space. The ShlA hemolysin of Serratia marcescens is synthesized as a proenzyme that contains an amino-terminal signal sequence, promoting its Secdependent translocation to the periplasm [166]. The ShlB polypeptide also contains an amino-terminal signal sequence and is exported through the Sec pathway [167]. Both ShlA and ShlB are processed by a signal peptidase structure and fold into mature species. The ShlB protein folds into a セM「。イ・ャ that inserts into the outer membrane. This event is required for the translocation of the enzymatic substrate ShIA. Other examples of this modified type V pathway include the secretion of filamentous hemagglutinin (FHA) by B. pertussis, and the secretion of the HpmA hemolysin by Proteus mirabilis [168].

Concluding Remarks

Molecular mechanisms that promote bacterial colonization are seemingly countless, however the accumulation of an ever-increasing body of information has allowed for the detection of common themes in pathogenesis. Strategies employed for the translocation of protein from the bacterial cytoplasm to targets in or beyond the cell wall envelope represent prime examples of th is commonality. Not only are the mechanisms for protein secretion conserved across species, many seemingly distinct secretion mechanisms share common components. Although protein secretion mechanisms represent only a fraction of the virulence strategies employed by bacteria, several of these processes represent primary virulence determinants. It appears that several pathogens use a combination of secretion mechanisms to establish infection, and the identification of mechanisms by analogy has allowed for rapid progress in the classification of a particular pathogen arsenal. This of course provides the potential for rapid biochemical characterization of secretion systems and their protein substrates, as weB as development and application of therapeutic targets to cover a wide range of bacterial species.

References

2 3

Navarre WW, Schneewind 0: Surface proteins of gram-positive bacteria and mechanisms of their largeling lo lhe cell wall envelope. Microbiol Mol BioI Rev 1999;63:174-229. Bruneteau M, Minka S: Lipopolysaccharides of bacterial pathogens from the genus Yersinia: A minireview. Biochimie 2003;85: 145-152. Berrington AW, Tan Ye, Srikhanta Y, Kuipers B, van der Ley P, Peak JR, Jennings MP: Phase variation in meningococcallipooligosaccharide biosynthesis genes. FEMS Immunol Med Microbiol 2002;34:267-275.

Bacterial 1nvasins

201

4 5 6 7 8 9 10 II 12 13 14 15

16 17 18 19 20 21

22

23

24

25

26

27 28

Inouye M: Bacterial Outer Membranes: Biogenesis and Functions. Wiley, New York, 1979. Missiakas D, Raina S: Protein folding in the bacterial periplasm. J Bacteriol 1997; 179:2465-2471. Driessen AJ, Fekkes P, Van Der Wolk JP: The Sec system. Curr Opin Microbiol 1998; 1:216-222. Danese PN, Silhavy TJ: Targeting and assembly of periplasmic and outer-membrane proteins in Escherichia coli. Annu Rev Genet 1998;32:59-94. Randall RR: Peptide binding by chaperone SecB: Implications for recognition of nonnative structure. Science 1992;257:241-245. Economou A, Wickner W: SecA promotes preprotein translocation by undergoing ATP-driven cycles of membrane insertion and deinsertion. Cell 1992;78:835-843. Dalbey RE, Lively MO, Bron S, Van Dijl JM: The chemistry and enzymology of the type I signal peptidases. Protein Sci 1997;6: 1129-1138. Lory S, Strom MS: Structure-function relationship of type-IV prepilin peptidase of Pseudomonas aeruginosa-A review. Gene 1997;192:117-121. Sankaran K, Wu HC: Bacterial prolipoprotein signal peptidase. Methods Enzymol 1995;248: 169-180. Poritz MA, Bernstein HD, Strub K., ZopfD, Wilhelm H, Walter P: An E. coli ribonucleoprotein containing 4.5S RNA resembles mammalian signal recognition particle. Science 1990;250: I II I-I I 17. Miller JD, Bernstein HD, Walter P: Interaction of E.coli Ffh/4.5S ribonucleoprotein and FtsY mimics that of mammalian signal recognition particle and its receptor. Nature 1994;367:657-659. Lingappa VR, Katz FN, Lodish HF, Blobel G: A signal sequence for the insertion ofa transmembrane glycoprotein. Similarities to the signals of secretory proteins in primary structure and function. J BioI Chern 1978;253:8667-8670. Davis NG, Model P: An artificial anchor domain: Hydrophobicity suffices to stop transfer. Cell 1985;41 :607--614. van Wely KHM, Swaving J, Freudl R, Driessen AJM: Translocation of proteins across the cell envelope of gram-positive bacteria. FEMS Microbiol Rev 2001;25:437--454. Stanley NR, Palmer T, Berks BC: The twin arginine consensus motif ofTat signal peptides is involved in Sec-independent protein targeting in Escherichia coli. J Bioi Chern 2000;275:11591-11596. Wu L-F, lze B, Chanal A, Quintin Y, Fichant G: Bacterial twin-arginine signal peptide-dependent protein translocation pathway: Evolution and mechanism. J Mol Microbiol Biotechnol 2000;2: 179-189. Santini C-L, Ize B, Chanal A, Muller M, Giordano G, Wu L-F: A novel Sec-independent periplasmic protein translocation pathway in Escherichia coli. EMBO J 1998;17: 101-112. Ochsner UA, Snyder A, Vasil AI, Vasil ML: Effects of the twin-arginine translocase on secretion of virulence factors, stress response, and pathogenesis. Proc Natl Acad Sci USA 2002;99: 8312-8317. Pradel N, Ye C, Livrelli V, Xu J, Joly B, Wu L-F: Contribution of the twin arginine translocation system to the virulence of enterohemorrhagic Escherichia coli 0 157:H7. Infect Immun 2003;71: 4908--4916. Bolhuis A, Mathers JE, Thomas ]D, Barrett CM, Robinson C: TatB and TalC form a functional and structural unit of the twin-arginine translocase complex of Escherichia coli. J Bioi Chern 200 I ;276:202 I3-202 I9. Sargent F, GoWke U, de Leeuw E, Stanley NR, Palmer T, Saibil HR, Berks BC: Purified components of the Escherichia coli Tat protein transport system form a double-layered ring structure. Eur J Biochem 2001;268:3361-3367. van Dijl JM, Braun PG, Robinson C, Quax WJ, Antelmann H, Hecker M, Muller J, Tjalsma H, Bron S, Jongbloed JDH: Functional genomic analysis of the Bacillus suMlis Tat pathway for protein secretion. J Biotechnol 2002;98:243-254. llangovan U, Ton-That H, Iwahara J, Schneewind 0, Clubb RT: Structure of sortase, the transpeptidase that anchors proteins to the cell wall of Staphylococcus aureus. Proc Natl Acad Sci USA 200 1;98:6056-6061. Mazmanian SK, Liu G, Ton-That H, Schneewind 0: Staphylococcus aureus sortase, an enzyme that anchors surface proteins to the cell wall. Science 1999;285:760-763. Mazmanian SK, Schneewind 0: Cell wall-anchored surface proteins and lipoproteins of gram-positive bacteria; in Sonenshein AL, Losick R, Hoch JA (eds): Bacillus sub/dis and Its Closest Relatives: From Genes to Cells. Washington, ASM Press, 2002.

Cambronne/Schneewind

202

29 30 31 32

33 34

35

36 37

38 39 40 41 42

43 44

45 46 47 48 49

50 51

Schneewind 0, Model P, Fischetti VA: Sorting of protein A to the staphylococcal cell wall. Cell 1992;70:267-281. 0: Cell wall sorting oflipoproteins in Staphylococcus aureus. Navarre WW, Daefler S, 、ョセ| ・ ョィ」s J BacterioI1996;178:441-446. Schneewind 0, Fowler A, Faull KF: Structure of the cell wall anchor of surface proteins in Staphylococcus aureus. Science 1995;268: I03-1 06. Mazmanian SK, Ton-That H, Su K, Schneewind 0: An iron-regulated sortase anchors a class of surface protein during Staphylococcus aureus pathogenesis. Proc Natl Acad Sci USA 2002;99: 2293-2298. Cossart P, Pizarro-Cerda J, Lecuit M: Invasion of mammalian cells by Listeria monocytogenes: Functional mimicry to subvert cellular functions. Trends Cell BioI 2003; 13:23-31. Bierne H, Mazmanian SK, Trost M, Pucciarelli G, Liu G, Dehoux P, Jiinsch L, Garcia-del Portillo F, Schneewind 0, Cossart P; European Listeria Genome Consortiwn: Inactivation of the srtA gene in Listeria monocytogenes inhibits anchoring of surface proteins and affects virulence. Mol Microbiol 2002;43:869-881. Jonquieres R, Fiedler P, Gounon P, Cossart P: Interaction between the protein JnlB of Listeria monocytogenes and lipoteichoic acid: A novel mechanism of protein association at the surface of gram-positive bacteria. Mol Microbiol 1999;34:902-914. Braun L, Ghebrehiwet B, Cossart P: gC Iq-Rlp32, a C Iq-binding protein, is a receptor for the InlB invasion protein of Listeria monocytogenes. EMBO J 2000; 19: 1458-1466. Mengaud J, Chenevert J, Geoffroy C, Gaillard JL, Cossart P: E-cadherin is the receptor for internalin, a surface protein required for entry of L. monocytogenes into epithelial cells. Cell 1996;84: 923-932. Shen Y, Naujokas M, Park M, Ireton K: InIB-dependent internalization of Listeria is mediated by the Met receptor tyrosine kinase. Cell 2000;103:501-510. Camilli A, Tilney LG, Portnoy DA: Dual roles of picA in Listeria monocytogenes pathogenesis. Mol MicrobioI1993;8:143-157. Gedde MM, Higgins DE, Tilney LG, Portnoy DA: Role of listeriolysin in cell-to-cell spread of Listeria monocytogenes. Infect Immun 2000;68:999-1003. Kocks C, Gouin E, Tabouret M, Berche P, Ohayon H, Cossart P: L. monocytogenes-induced actin assembly requires the aclA gene product, a surface protein. Cell 1992;68:521-531. Pistor S, Chakaraborty T, Niebuhr K, Domann E, Wehland J: The ActA protein of Listeria monocytogenes acts as a nucleator inducing reorganization of the actin cytoskeleton. EMBO J 1994; 13:758-763. Bear JE, Krause M, Gertler FB: Regulating cellular actin assembly. Cun Opin Cell Bioi 2001;13: 158-166. Marquis H, Goldfine H, Portnoy DA: Proteolytic pathways of activation and degradation ofa bacterial phospholipase C during intracellular infection by Listeria monocytogenes. J Cell Bioi 1997; 137: 1381-1392. Isberg RR, Voorhis DL, Falkow S: Identification of invasin: A protein that allows enteric bacteria to penetrate cultured mammalian cells. Cell 1987;50:769-778. Hamburger ZA, Brown MS, Jsberg RR, Bjorkman PJ: Crystal structure of invasin: A bacterial integrin-binding protein. Science 1999;286:291-295. Leong 1M, Fournier RS, Isberg RR: Mapping and topographic localization of epitopes of the Yersinia pseudotuberculosis invasin protein. Infect Immun 1991 ;59:3424-3433. Isberg RR, Leong JM: Multiple 13-1 chain integrins are receptors for invasin, a protein that promotes bacterial penetration into mammalian cells. Cell 1990;60:861-871. Schulte R, Kerneis S, Klinke S, Bartels H, Preger S, Kraehenbuh1 J-p, Pringault E, Autenrieth IB: Translocation of Yersinia enterocolitica across reconstituted intestinal epithelial monolayers is triggered by Yersinia invasin binding to 13-1 integrins apically expressed on M-like cells. Cell Microbiol 2000;2: 173-185. Lindberg P, Lund B, Johansson L, Normark S: Localization of the receptor-binding protein adhesin at the tip of the bacterial pilus. Nature 1987;328:84-87. Pasloske BL, Finlay BB, Paranchych W: Cloning and sequencing of the Pseudomonas aeruginosa PAK pilin gene. FEBS Lett 1985;183:408-412.

°

Bacterial Invasins

203

52 53 54

55 56 57 58

59 60

61

62

63 64

65 66 67

68 69 70

71 72

73

74

Koronakis V, Hughes C: Synthesis, maturation and export of the E. coli hemolysin. Med Microbiol ImmunoI1996;185:65-71. Fath MJ, Kolter R: ABC transporters: Bacterial exporters. Microbiol Rev 1993;57:995-1017. Wang RC, Seror SJ, Blight M, Pratt JM, Broome-Smith JK, Holland IB: Analysis of the membrane organization of an Escherichia coli protein trans locator, HlyB, a member of a large family of prokaryote and eukaryote surface transport proteins. J Mol Bioi 199 1;217:441-454. Stanley P, Packman LC, Koronakis V, Hughes C: Fatty acylation of two internal lysine residues required for the toxic activity of Escherichia coli hemolysin. Science 1994;266: I992-1 996. Issartel Jp, Koronakis V, Hughes C: Activation of Escherichia coli prohaemolysin to the mature toxin by acyl carrier protein-dependent fatty acylation. Nature \99\ ;35\ :759-76\. Ludwig A, Vogel M, Goebel W: Mutations affecting activity and transport of haemolysin in Escherichia coli. Mol Gen Genet 1987;206:238-245. Thanabalu T, Koronakis E, Hughes C, Koronakis V: Substrate-induced assembly of a contiguous channel for protein export from E. coli: Reversible bridging of an inner-membrane translocase to an outer membrane exit pore. EMBO J 1998; I7:6487--6496. Gray L, Baker K, Kenny B, Mackman N, Haigh R, Holland IB: A novel C-terminal signal sequence targets E. coli haemolysin directly to the medium. Mol Gen Genet 1989;205: 127-133. Koronakis V, Koronakis E, Hughes C: Isolation and analysis of the C-terminal signal directing export of Escherichia coli hemolysin protein across both bacterial membranes. EMBO J 1989;8: 595--605. Koronakis V, Sharf[ A, Koronakis E, Luisi B, Hughes C: Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export. Nature 2000;405: 914-919. Koronakis V, Hughes C, Koronakis E: ATPase activity and ATP/ADP-induced conformational change in the soluble domain of the bacterial protein translocator HlyB. Mol Microbiol 1993;8: 1163- 1175. Felmlee T, Welch RA: Alterations of amino acid repeats in the Escherichia coli hemolysin affect cytolytic activity and secretion. Proc Natl Acad Sci USA 1988;85:5269-5273. Stathopoulos C, Hendrixson DR, Thanassi DG, Hultgren SJ, St Geme JW 3rd, Curtiss R 3rd: Secretion of virulence determinants by the general secretory pathway in gram-negative pathogens: An evolving story. Microbes Infect 2000;2:1061-1072. Russel M: Filamentous phage assembly. Mol Microbiol 1991 ;5: 1607-1613. Bender H, Wallenfels K: Pullulanase (an amylopectin and glycogen debranching enzyme) from Aerobacter aerogenes. Methods Enzymol \966;8:555-559. Pugsley Ap, Francetic 0, Possot OM, Sauvonnet N, Hardie KR: Recent progress and future directions in studies of the main terminal branch of the general secretory pathway in gram-negative bacteria - A review. Gene 1997; 192: I 3- I9. Pugsley AP, Chapon C, Schwartz M: Extracellular pullulanase of Klebsiella pneumoniae is a lipoprotein. J Bacteriol 1986; 166: I083- 1088. Possot OM, Pugsley AP: The conserved tetracysteine motif in the general secretory pathway component PulE is required for efficient pullulanase secretion. Gene 1997; I92:45-50. Sandkvist M, Bagdasarian M, Howard SP, DiRita VJ: Interaction between the autokinase EpsE and EpsL in the cytoplasmic membrane is required for extracellular secretion in Vibrio cholerae. EMBO J 1995;14:1664-1673. Russel M: Macromolecular assembly and secretion across the bacterial cell envelope: Type II protein secretion systems. J Mol Bioi \998;279:485-499. Pugsley AP, Dupuy B: An enzyme with type IV prepilin peptidase activity is required to process components of the general extracellular protein secretion pathway of Klebsiella oxytoca. Mol Microbiol 1992;6:751-760. Hobbs M, Mattick JS: Common components in the assembly of type 4 fimbriae, DNA transfer systems, filamentous phage and protein-secretion apparatus: A general system for the formation of surface-associated protein complexes. Mol Microbiol 1993; 10:233-243. Nouwen N, Ranson N, Saibil H, Wolpensinger B, Engel A, Ghazi A, Pugsley AP: Secretin PuID: Association with pilot PuiS, structure, and ion-conducting channel formation. Proc Natl Acad Sci USA 1999;96:8173-8177.

Cambronne/Schneewind

204

75 76 77

78 79 80 81 82 83 84

85 86

87 88

89 90 91

92 93

94

95

96 97 98

Schmitt CK, Meysick KC, 0 'Brien AD: Bacterial toxins: Friends or foes? Emerg Infect Dis 1999; 5:224-234. Lonnroth I, Holmgren J: Subunit structure of cholera toxin. J Gen Microbiol 1973;76:417--427. Sandkvist M, Michel LO, Hough LP, Morales VM, Bagdasarian M, Koomey M, DiRita VJ: General secretion pathway (eps) genes required for toxin secretion and outer membrane biogenesis in Vibrio cholerae. J Bacteriol 1997; 179:6994-7003. Cuatrecasas P: Interaction of Vibrio cholerae enterotoxin with cell membranes. Biochemistry 1973; 12:3547-3558. Cassel D, Selinger Z: Mechanism of adenyl ate cyclase activation by cholera toxin: Inhibition of GTP hydrolysis at the regulatory site. Proc Natl Acad Sci USA 1977;74:3307-3311. Rosqvist R, Magnusson K-E, Wolf-Watz H: Target cell contact triggers expression and polarized transfer of Yersinia YopE cytotoxin into mammalian cells. EMBO J 1994; 13:964-972. Hueck CJ: Type III protein secretion in bacterial pathogens of animals and plants. Microbiol Mol BioI Rev 1998;62:379--433. Winstanley C, Hart CA: Type 111 secretion systems and pathogenicity islands. J Med Microbiol 2001 ;50: 116-126. Pallen MJ, Chaudhuri RR, Henderson IR: Genomic analysis of secretion systems. CUff Opin Microbiol 2003;6:519-527. Young GM, Schmiel DH, MillerVL: A new pathway for the secretion of virulence factors by bacteria: The flagellar export apparatus functions as a protein secretion system. Proc Natl Acad Sci USA 1999;96:6456-6461. Cornelis GR, Boland A, Boyd AP, Geuijen C, Iriarte M, Neyt C, Sory M-P, Stainier I: The virulence plasmid of Yersinia, an antihost genome. Microbiol Mol Bioi Rev 1998;62: 1315-1352. Allaoui A, Schulte R, Comelis GR: Mutational analysis of the Yersinia enterocolitica virC operon: Characterization of yscE, F, G, I, J, K required for Yop secretion and yscH encoding YopR. Mol Microbiol 1995; 18:343-355. Allaoui A, Woestyn S, Sluiters C, Comelis GR: YscU, a Yersinia enterocolitica inner membrane protein involved in Yop secretion. J Bacteriol 1994; 176:4534--4542. Bergmann T, Hakansson S, Forsberg A, Norlander L, MaceliaroA, Backman A, Bolin I, Wolf-Watz H: Analysis of V antigen IcrGVH-yopBD operon of Yersinia pseudotuberculosis: Evidence for a regulatory role of IcrH and IcrV. J Bacteriol 1991; 173: 1607-1616. Macnab RM: How bacteria assemble flagella. Annu Rev MicrobioI2003;57:77-100. Kimbrough TG, Miller SI: Contribution of Salmonella typhimurium type III secretion components to needle complex formation. Proc Natl Acad Sci USA 2000;97: I 1008-1 1013. Michiels T, Vanooteghem J-C, Lambert de Rouvroit C, China B, Gustin A, Boudry P, Comelis GR: Analysis of virC, an operon involved in the secretion ofYop proteins by Yersinia enterocolitica. J Bacteriol 1991; 173:4994-5009. Woestyn S, Allaoui A, Wattiau P, Cornel is GR: YscN, the putative energizer of the Yersinia Yop secretion machinery. J Bacteriol 1994; 176: 1561-1569. Genin S, Boucher CA: A superfamily of proteins involved in different secretion pathways in gramnegative bacteria: Modular structure and specificity of the N-terminal domain. Mol Gen Genet 1994;243:112-118. Koster M, Bitter W, de Cock H, A1laoui A, Comelis GR, Tommassen J: The outer membrane component, YscC, of the Yop secretion machinery of Yersinia enterocolitica forms a ring-shaped multimeric complex. Mol Microbiol 1997;26:789-798. Edqvist PJ, Olsson J, Lavander M, Sundberg L, Forsberg A, Wolf-Watz H, Lloyd SA: YscP and YscU regulate substrate specificity of the Yersinia type III secretion system. J Bacteriol 2003; 185:2259-2266. Hoiczyk E, Blobel G: Polymerization of a single protein of the pathogen Yersinia enterocolitica into needles punctures eukaryotic cells. Proc Nail Acad Sci USA 200 I ;98:4669--4674. Payne PL, Straley SC: YscO of Yersinia pestis is a mobile core component of the Yop secretion system. J Bacteriol 1998; 180:3882-3890. Kubori T, Shimamoto N, Nakamura D, Uralil J, Lara-Tejero M, Sukban A, Galan JE, Aizawa SI: Supramolecular structure of the Salmonella typhimurium type III protein secretion system. Science 1998;280:602-605.

Bacterial Invasins

205

99

100

10 I

102

103 104

105

106 107 108 109 I 10 III

I 12 113 114 115

116

117

118 119

120

Sekiya K, Ohishi M, Ogino T, Tamano K, Sasakawa C, Abe A: Supermolecular structure of the enteropathogenic Escherichia coli type III secretion system and its direct interaction with the EspA-sheath-like structure. Proc Natl Acad Sci USA 200 I;98: I 1638-1 1643. Tamano K, Aizawa S, Katayama E, Nonaka T, Imajoh-Ohmi S, Kuwae A, Nagai S, Sasakawa C: Supramolecular structure of the Shigella type III secretion machinery: The needle part is changeable in length and essential for delivery of effectors. EMBO J 2000; 19:3876-3887. Blocker A, Jouihri N, Larquet E, Gounon P, Ebel F, Parsot C, Sansonetti P, Allaoui A: Structure and composition of the Shigellajlexneri 'needle complex', a part of its type III secretion. Mol Microbiol 2001 ;39:652-663. Cordes FS, Komoriya K, Larquet E, Yang S, Egelman EH, Blocker A, Lea SM: Helical structure of the needle of the type III secretion system of Shigella jlexneri. J Bioi Chern 2003;278: 17103-17107. Hakansson S, Bergman T, Vanooteghem J-C, Comelis G, Wolf-Watz H: YopB and YopD constitute a novel class of Yersinia Yop proteins. Infect Immun 1993;61:71-80. Neyt C, Comelis GR: Insertion of a Yop translocation pore into the macrophage plasma membrane by Yersinia enterocolitica: Requirement for trans locators YopB and YopD, but not LcrG. Mol Microbiol 1999;33:971-981. Persson C, Nordfelth R, Holmstrom A, Hakansson S, Rosqvist R, Wolf-Watz H: Cell-surfacebound Yersinia trans locate the protein tyrosine phosphatase YopH by a polarized mechanism into the target cell. Mol MicrobioI1995;18:135-150. Sarker MR, Neyt C, Stainier I, Comelis GR: The Yersinia yop virulon: LcrV is required for extrusion of the translocators YopB and YopD. J Bacteriol 1998;180: 1207-1214. Anderson OM, Schneewind 0: A mRNA signal for the type III secretion of Yop proteins by Yersinia enterocolitica. Science 1997;278:1140-1143. Anderson OM, Schneewind 0: Yersinia enterocolitica type III secretion: An mRNA signal that couples translation and secretion ofYopQ. Mol MicrobioI1999;31:1139-1148. Lloyd SA, Norman M, Rosqvist R, Wolf-Watz H: Yersinia YopE is targeted for type III secretion by N-terminal, not mRNA, signals. Mol Microbiol 2001 ;39:520-531. Ramamurthi KS, Schneewind 0: Yersinia enterocolitica type III secretion: Mutational analysis of the yopQ secretion signal. J Bacteriol 2002; 184:3321-3328. Sory M-P, Boland A, Lambermont I, Comelis GR: Identification of the YopE and YopH domains required for secretion and internalization into the cytosol of macrophages, using the cyaA gene fusion approach. Proc Natl Acad Sci USA 1995;92: I 1998-12002. Birtalan SC, Phillips RM, Ghosh P: Three-dimensional secretion signals in chaperone-effector complexes of bacterial pathogens. Mol Cell 2002;9:971-980. Cheng LW, Anderson OM, Schneewind 0: Two independent type III secretion mechanisms for YopE in Yersinia enterocolitica. Mol Microbiol 1997;24:757-765. Black OS, Bliska 18: The RhoGAP activity of the Yersinia pseudotuberculosis cytotoxin YopE is required for antiphagocytic function and virulence. Mol Microbiol 2000;37:515-527. Von Pawal-Rammingen U, Telepnev MY, Schmidt G, Aktories K, Wolf-Watz H, Rosqvist R: GAP activity of the Yersinia YopE cytotoxin specifically targets the Rho pathway: A mechanism for disruption of actin microfilament structure. Mol Microbiol 2000;36:737-748. Zhang ZY, Clemens JC, Schubert HL, Stuckey JA, Fischer MW, Hume OM, Saper MA, Dixon JE: Expression, purification, and physicochemical characterization of a recombinant Yersinia protein tyrosine phosphatase. J Bioi Chern 1992;267:23759-23766. Deleuil F, Mogemark L, Francis MS, Wolf-Watz H, Fallman M: Interaction between theYersinia protein tyrosine phosphatase YopH and eukaryotic Cas/Fyb is an important virulence mechanism. Cell Microbiol 2003;5:53-64. Hamid N, Gustavsson A, Andersson K, McGee K, Persson C, Rudd CE, Fallmen M: YopH dephosphorylates Cas and Fyn-binding protein in macrophages. Microb Pathog 1999;27:231-242. Sauvonnet N, Predet-Balade B, Garcia-Sanz JA, Cornelis GR: Regulation ofmRNA expression in macrophages following Yersinia enterocolitica infection: Role of different Yop effectors. J Bioi Chern 2002;227:25133-25142. Skrzypek E, Cowan C, Straley SC: Targeting of the Yersinia pestis YopM protein into He La cells and intracellular trafficking to the nucleus. Mol Microbiol 1998;30: 1051-1065.

Cambronne/Schneewind

206

121

122 123 124 125

126

127

128

129

130 131 132 133 134

135

136 137 138 139 140

141

142

Juris SJ, Rudolph AE, Huddler 0, Orth K, Dixon JE: A distinctive role for the Yersinia protein kinase: Actin binding, kinase activation, and cytoskeleton disruption. Proc Nat! Acad Sci USA 2000;97:9431-9436. Galyov EE, Hakansson S, Forsberg A, Wolf-Watz H: A secreted protein kinase of Yersinia pseudotuberculosis is an indispensable virulence determinant. Nature 1993;361:730-732. Boland A, Comelis GR: Role of YopP in suppression of tumor necrosis factor alpha release by macrophages during Yersinia infection. Infect Immun 1998;66: 1878-1884. Orth K, Palmer LE, Bao ZQ, Stewart S, Rudolph AE, Bliska 18, Dixon JE: Inhibition of the mitogenactivated protein kinase kinase superfamily by a Yersinia effector. Science 1999;285: 1920-1923. Schesser K, Spiik AK, Dukuzumuremyi JM, Neurath MF, Pettersson S, Wolf-Watz H: TheyopJ locus is required for Yersinia-mediated inhibition of NF-kappaB activation and cytokine expression: YopJ contains a eukaryotic SH2-like domain that is essential for its repressive activity. Mol Microbiol 1998;28: 1067-1079. Orth K, Xu Z, Mudgett MB, Bao ZQ, Palmer LE, Bliska 18, Mangel WF, Staskawicz B, Dixon JE: Disruption of signaling by Yersinia effector YopJ, a ubiquitin-like protein protease. Science 2000;290: 1594-1597. Denecker G, Declercq W, Geuijen CA, Boland A, Benabdillah R, van Gurp M, Sory M-P, Vandenabeele P, Cornel is GR: Yersinia enlerocolilica YopP-induced apoptosis of macrophages involves the apoptotic signaling cascade upstream of bid. J Bioi Chern 200 1;276: 19706-19714. Shao F, Merritt PM, Bao Z, Innes RW, Dixon JE: A Yersinia effector and a Pseudomonas avirulence protein define a family of cysteine proteases functioning in bacterial pathogenesis. Cell 2002;109:575-588. Jarvis KG, Giron JA, Jerse AE, McDaniel TK, Donnenberg MS, Kaper 18: Enteropathogenic Escherichia coli contains a putative type III secretion system necessary for the export of proteins involved in attaching and effacing lesion formation. Proc Natl Acad Sci USA 1995;92:7996-8000. Knutton S, Lloyd DR, McNiesh AS: Adhesion of enteropathogenic Escherichia coli to human intestinal enterocytes and cultured human intestinal mucosa. Infect Immun 1987;57:69-77. Kenny B, DeVinney R, Stein M, Reinscheid OJ, Frey EA, Finlay BB: Enteropathogenic E coli (EPEe) transfers its receptor for intimate adherence into mammalian cells. Cell 1997;91 :511-520. Hueck CJ, Hantman MJ, Bajaj V, Johnston C, Lee CA, Miller SI: Salmonella typhimurium secreted invasion determinants are homologous to Shigella Ipa proteins. Mol MicrobioI1995;18:479-490. Zhou 0, Mooseker MS, Galan JE: Role of the S. typhimurium actin-binding protein SipA in bacterial internalization. Science 1999;283:2092-2095. Hardt WD, Chen LM, Schuebel KE, Bustelo WR, Galan JE: Salmonella typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 1998;93:815-826. Stender S, Friebel A, Linder S, Rhode M, Mirold S, Hardt WD: Identification of SopE2 from Salmonella typhimurium, a conserved guanine nucleotide exchange factor for Cdc42 of the host cell. Mol Microbiol 2000;36: 1206-1221. Fu Y, Galan JE: A Salmonella protein antagonizes Rac-I and Cdc42 to mediate host-cell recovery after bacterial invasion. Nature 1999;40 I:293-297. Fu Y, Galan JE: The Salmonella typhimurium tyrosine phosphatase SptP is translocated into host cells and disrupts the actin cytoskeleton. Mol Microbiol 1998;27:359-368. Hensel M: Salmonella pathogenicity island 2. Mol Microbiol 2000;36: 1015-1023. Zychlinsky A, Prevost MC, Sansonetti PJ: Shigella flexneri induces apoptosis in infected macrophages. Nature 1992;358: 167-169. Sansonetti PJ: Microbes and Microbial Toxins: Paradigms for Microbial-Mucosal Interactions Ill. Shigellosis: From symptoms to molecular pathogenesis. Am J Physiol Gastrointest Liver Physiol 2001 ;280:319-323. Bernardini ML, Mounier J, d'Hautevilie H, Coquis-Rondon M, Sansonetti PJ: Identification of icsA, a plasmid locus of Shigella flexneri which governs bacterial intra- and intercellular spread through interaction with F-actin. Proc Natl Acad Sci USA 1989;86:3867-3871. Egile C, Loisel Tp, Laurent V, Rong Li Pantaloni 0, Sansonetti PJ, Carlier MF: Activation of the CDC42 effector N- WASP by the Shigella icsA protein promotes actin nucleation by Arp2/3 complex and bacterial actin-based motility. J Cell Bioi 1999;146:1319-1332.

Bacterial Invasins

207

143 144 145

146 147 148 149 150 151

152

153 154

155 156 157 158 159 160

161 162 163

164 165

166

Burns DL: Type IV transporters of pathogenic bacteria. Curr Opin MicrobioI2003;6:29-34. Cascales E, Christie PJ: The versatile bacterial type IV secretion systems. Nat Rev Microbiol 2003;1: 137-149. Nicosia A, Perugini M, Franzini C, Casagli MC, Borri MG, Antoni G, Almoni M, Neri P, Ratti G, Rappuoli R: Cloning and sequencing of the pertussis toxin genes: Operon structure and gene duplication. Proc Natl Acad Sci USA 1986;83:4631-4635. Pansegrau W, Lanka E: Enzymology of DNA transfer by conjugative mechanisms. Prog Nucleic Acid Res Mol Bioi 1996;54: 197-251. Kuldau GA, De Vos G, Owen J, McCaffrey G, Zambryski P: The virB operon of Agrobacterium tumejaciens pTiC58 encodes II open reading frames. Mol Gen Genet 1990;221 :256-266. Lai EM, Kado Cl: Processed VirB2 is the major subunit of the promiscuous pilus of Agrobacterium twnejaciens. J Bacteriol 1998; 180:2711-2717. Anderson LB, Hertzel AY, Das A: Agrobacterium tumejaciens VirB7 and VirB9 form a disulfidelinked protein complex. Proc Natl Acad Sci USA 1996;93:8889-8894. Das A, Xie YH: Construction of transposon Tn3phoA: Its application in defining the membrane topology of the Agrobacteriwn twnejaciens DNA transfer proteins. Mol Microbiol 1998;27:405-414. Kumar RB, Xie YH, Das A: Subcellular localization of the Agrobaclerium tume(aciens T-DNA transport pore proteins: VirB8 is essential for the assembly of the transport pore. Mol Microbiol 2000;36:608-617. Savvides SN, Yeo HJ, Beck MR, Blaesing F, Lurx R, Lanka E, Buhrdorf R, Fischer W, Haas R, Waksman G: VirBll ATPases are dynamic hexameric assemblies: New insights into bacterial type IV secretion. EMBO J 2003;22: 1969-1980. Yeo H-J, Savvides SN, Herr AB, Lanka E, Waksman G: Crystal structure of the hexameric traffic ATPase of the Helicobacter pylori type IV secretion system. Moll Cell 2000;6: 1461-1472. Farizo KM, Huang T, Burns DL: Membrane localization of the S I subunit of pertussis toxin in Bordetella pertussis and implications for pertussis toxin secretion. Infect Immun 2002;70: 1193-1201. Covacci A, Telford JL, Del Giudice G, Parsonnet J, Rappuoli R: Helicobacter pylori virulence and genetic geography. Science 1999;284: 1328-1333. Odenbreit S, PuIs J, Sedlmaier B, Gerland E, Fischer W, Haas R: Translocation of Helicobacter pylori CagA into gastric epithelial cells by type IV secretion. Science 2000;287: 1497-1500. Roy CR: Exploitation of the endoplasmic reticulum by bacterial pathogens. Trends Microbiol 2002; I0:418-424. Vogel JP, Andrews HL, Wong SK, lsberg RR: Conjugative transfer by the virulence system of Legionella pnewnophila. Science 1998;279:873-876. Nagai H, Kagan JC, Zhu X, Kahn RA, Roy CR: A bacterial guanine nucleotide exchange factor activates ARF on Legionella phagosomes. Science 2002;295:679--682. Conover GM, Derre I, Vogel JP, lsberg RR: The Legionella pheumophila LidA protein: A translocated substrate of the DotJIcm system associated with maintenance of bacterial integrity. Mol Microbiol 2003;48:305-321. Halter R, Pohlner J, Meyer TF: IgA protease of Neisseria gonorrhoeae: Isolation and characterization of the gene and its extracellular product. EMBO J 1984;3:1595-1601. Pohlner J, Halter R, Beyreuther K, Meyer TF: Gene structure and extracellular secretion of Neisseria gonorrhoeae IgA protease. Nature 1987;325:458-462. Klauser T, Pohlner J, Meyer TF: Extracellular transport of cholera toxin B subunit using Neisseria IgA protease domain: Conformation-dependent outer membrane translocation. EM BO J 1990;9: 1991-1999. Egile C, D'Hauteville H, Parsot C, Sansonetti PJ: SopA, the outer membrane protease responsible for polar localization of IcsA in Shigellajlexneri. Mol Microbiol 1997;23: I063-1 073. Roggenkamp A, Ackermann N, Jacobi CA, Truelzsch K, Hoffmann H, Heesemann J: Molecular analysis of transport and oligomerization of the Yersinia enterocolitica adhesin YadA. J Bacteriol 2003; 185:3735-3744. Schonherr R, Tsolis R, Focareta T, Braun V: Amino acid replacements in the Serratia marcescens haemolysin ShlA define sites involved in activation and secretion. Mol Microbiol 1993;9: 1229-1237.

Cambronne/Schneewind

208

167

168

Konninger Uw, Hobbie S, Benz R, Braun V: The haemolysin-secreting ShlB protein of the outer membrane of Serratia marcescens: Determination of surface-exposed residues and formation of ion-permeable pores by ShlB mutants in artificial lipid bilayer membranes. Mol Microbiol 1999;32: 1212-1225. Jacob-Dubuisson F, Buisine C, Willery E, Renauld-Mongenie G, Locht C: Lack of functional complementation between Bordetella pertussis filamentous hemagglutinin and Proteus mirabilis HpmA hemolysin secretion machineries. J BacterioI1997;179:775-783.

Eric D. Cambronne Section of Microbial Pathogenesis, Yale University School of Medicine 295 Congress Avenue, New Haven, CT 06536 (USA) Tel. + 1 203 7372404, E-Mail [email protected]

Bacterial Invasins

209

Signaling and Gene Regulation Russell W, HelWald H (eds): Concepts in Bacterial Virulence. Contrib Microbiol. Basel, Karger, 2005, vol 12, pp 210-233

Bacterial Iron Transport Related to Virulence Volkmar Braun MikrobiologielMembranphysiologie, Universitat Tiibingen, Tiibingen, Germany

The Problem of Iron Supply

Under oxic conditions, iron occurs in the Fe3+ valence state and forms insoluble polymeric hydroxyl-aquo complexes. Therefore, all aerobically living organisms that contain iron in many cytosolic and membrane-bound redox proteins, in particular in respiratory chains, have developed means to solubilize Fe3+. Bacteria and fungi synthesize iron-complexing compounds, designated siderophores, which are secreted, bind extracellular Fe3+, and are transported as Fe3+ complexes via specific transport systems into the cells, where Fe3+ is released from the complexes, usually by reduction to Fe 2 +, and then incorporated into heme, iron-sulfur proteins, and other forms of protein reaction centers. Higher organisms synthesize heme, which is the most abundant form of iron-containing compounds. Only a small percentage of the heme occurs in free form; most of it is incorporated into hemoglobin and bound to hemopexin. Important extracellular iron-binding proteins in higher organisms are lransferrin and lacloferrin and inlracellular ferrilin. Transferrin is lhe predominant iron carrier that delivers iron to cells. The di-iron complex is taken up by transferrin receptors, and the iron is released in endosomes and then further metabolized. Lactoferrin is the predominant iron-binding protein in secretory fluids. Transferrin and lactoferrin bind Fe 3 + so tightly that the free Fe 3 + concentration in equilibrium with these proteins is in the order of I ion per liter. The extreme lack of iron inhibits growth of microorganisms. However, some bacteria synthesize transferrin and lactoferrin receptor proteins exposed at the bacterial cell surfaces, which remove the iron from transferrin and lactoferrin and transport iron across the outer membrane.

This short overview focuses on some prominent examples of iron supply systems formed by human pathogenic bacteria. The reader is referred to more comprehensive reviews on specific aspects [1-24].

Overview of Bacterial Iron Transport Systems

Transport across the Cytoplasmic Membrane The design of Fe3+ transport systems across the cytoplasmic membrane is the same for gram-negative and gram-positive bacteria. The systems belong to the ATP-binding cassette (ABC) transporters, which consist of a binding protein, a permease, and an ATPase (fig. 1). The binding proteins of gram-negative bacteria are located in the periplasm. In gram-positive bacteria, the binding proteins are linked by a lipid of the murein-lipoprotein type (triacyl-glyceryl cysteine) to the outer surface of the cytoplasmic membrane. The permease consists of one or two proteins that are incorporated into the cytoplasmic membrane and translocate Fe3+, Fe3+-siderophores, or heme across the cytoplasmic membrane. The ATPase provides the energy derived from ATP binding and subsequent ATP hydrolysis [25]. Crystal structures have been determined for two Fe3+-binding proteins, FbpA of Neisseria gonorrhoeae and hFbpA of Haemophilus influenzae [26], and for the ferrichrome-binding protein FhuD, which binds structurally related siderophores of the hydroxamate type and the antibiotic albomycin [16, 27]. The crystal structures of FbpA and hFbpA are similar, but differ from that of FhuD. The three proteins are composed of two globular domains; in FbpA and hFbpA, these domains are connected by a hinge region that permits closure of the globular domains upon binding of Fe3+ (like a Venus fly trap). In contrast, the two globular domains of FhuD are connected by a rigid, kinked a-helix that allows only a slight movement of the globular domains. The crystal structure of an entire ABC transporter, the vitamin B 12 transporter of Escherichia coli, has recently been unraveled. The ABC transporter consists of the BtuC permease and associated BtuD ATPase [28], and the BtuF-binding protein [29]. Since the BtuF structure is similar to FhuD and the transmembrane topology of BtuC is comparable to that of FhuB [15] which transports ferrichrome across the cytoplasmic membrane [30], it is predicted that the structure of the vitamin B 12 transport system is representative for the ferric siderophore and heme transport systems. BtuF can he positioned via salt bridges on top of the BtuC permease. BtuCD forms a translocation channel that is large enough to accommodate vitamin B 12 • In the crystal, the channel is open to the periplasmic side and closed to the cytoplasmic side. BtuD controls opening of the BtuC channel. The two BtuD subunits located at the inner side ofthe

Iron Transport Related to Virulence

211

OM

PP

o

01------

.D

.D

W

tl:l

x

FhuC

()

1

a

ATP

c

Albomycin

Braun

ADP+P j

CM

FhuC

() ATP

b

ADP+P j

Rifamycin CGP4832

d

212

-:Y

hセキ セ

I

NH セ

1\

O-+-CO 0 '---/ o

N

0

CH 3

1

e

Albomycin

f

Rifamycin CGP4832

Fig. 1. Crystal structLUe of the FhuA outer membrane (OM) transport protein of E. coli with bound antibiotics albomycin (a) and rifamycin (b) CGP 4832, which are transported by FhuA. The structures of the antibiotics derived from tbe crystal structures (c, d) and the chemical formula (e,/) are shown. a, b The model illustrates the subcellular location of the proteins TonB, ExbB, and ExbD, which form the energy-transducing complex between the cytoplasmic membrane and the outer membrane, the transport proteins across the cytoplasmic membrane, and the interactions of the proteins. This protein arrangement is typical for all transport systems of gram-negative bacteria that transport Fe3+, Fe 3 + -siderophores, and berne. For further information, see the text. PP = Periplasm; CM = cytoplasmic membrane.

cytoplasmic membrane are in close contact to the two BtuC subunits. Binding of ATP moves the two BtuD subunits closer together. This might rearrange the two BtuC subunits such that the channel opens to the cytoplasmic side. BtuF loaded with vitamin B t2 is bound to BtuC, delivers vitamin B 12 to BtuC, and triggers ATP hydrolysis. The BtuD molecules move apart, which in tum closes the BtuC channel to the cytoplasmic side and opens it to the periplasmic side for the next round of vitamin B 12 transport.

Transport across the Outer Membrane Gram-negative bacteria contain an outer membrane that forms a permeability barrier for hydrophilic substrates above a certain molar mass, which in E. coli is 600 daltons [31]. The inner diameter of the porins through which the substrates diffuse across the outer membrane determines the substrate size. The Fe3+ siderophores usually have a molecular weight greater than 600 and cannot

Iron Transport Related to Virulence

213

diffuse with a sufficient rate through porins. In addition, their concentration is too low for diffusion to satisfy the growth requirement - in the order of 10 5 iron ions per cell per generation. The siderophores, heme, and the iron-binding proteins adsorb to outer membrane proteins, which not only serve as receptors but also function as transporters across the outer membrane. The iron compounds are thereby concentrated at the bacterial cell surface and are subsequently actively transported by an energy-consumjng process across the outer membrane into the periplasm. There is no energy source in the outer membrane to drive active transport. Energy is provided by the cytoplasmic membrane through the proton motive force [32]. TonB, ExbB, and ExbD are the three known proteins that relay the energy from the cytoplasmic membrane into the outer membrane [33, 34]. These proteins are located in the cytoplasmjc membrane and interact with each other, and TonE interacts with the outer membrane transport proteins. It is thought that these three proteins respond to the proton motive force of the cytoplasmic membrane (e.g., the proton gradient), react with a conformational change, and store the energy as potential energy. Upon interaction of energized TonB with the outer membrane transporters, the bound iron compounds are released from their binding sites and a channel is opened through which the iron compounds diffuse into the periplasm. The crystal stmctures of three outer membrane iron transporters FhuA [35, 36], FepA [37], and FecA [38, 39], and the vitamin B l2 transporter BtuB [40] provide a conceptual framework of how these transporters might function. The structures reveal a セM「。イ・ャ composed of 22 antiparallel セMウエイ。ョ、 that form a channel. The channel is closed by a globular domain, which is designated as the cork plug, or hatch. Binding of the substrates to the transporters occurs at a site well above the cell surface. Very strong binding occurs through approximately ten-amino aCi1 side chains with a binding constant in the nanomolar range. Energy input is required to release the substrates from their binillng sites and to move the cork so that a channel is formed through which the substrates gain access to the periplasm. The theory is that TonB transfers potential energy to the transporters, which alte I their confonnation to open a chalmeJ. TonB is deenergized, and the transporters close the channels after the iron compounds have passed through by diffusion. The genetically and biocherrucally identified sites of interaction between TonE and the transporters are located in the TonE box of the transporters and a region around residue 160 of TonE [41, 42]. The crystal structures and electron spin resonance deterrrunations of nitroxide-substituted TonB box residues of BtuB demonstrate that the TonB box is exposed to the periplasm and moves upon binding of the substrates to the transporters [43]. The TonB box and the substrate-binding sites are far apart, wrnch implies long-range stmctural transitions throughout the entire transporter. Transport across the outer membrane is mechanistically not coupled

Braun

214

to transport across the cytoplasmjc membrane. The two membrane transport processes occur independently of each other.

Iron Transport Associated with Virulence

Iron-Controlled Bacterial Functions Since iron is an essential element, but available only in growth-limiting concentrations, those bacteria that multiply in the human body express potent iron transport systems. The relationship of iron transport to virulence is usually not easy to establish since bacteria normally express several iron transport systems. Knocking out one system by mutation might not result in conversion of a pathogenjc strain to a nonpathogenic strain since other iron transport systems take over the iron supply. For example, a pathogenic E. coli strain may transport Fe H by the siderophores aerobactin, enterobactin, salmochelin, citrate, ferrichrome, and heme, and Fe2+ via thefeo-encoded transport system. tonB, exbB, and exbD are the only genes involved in all energy-coupled outer membrane iron transport systems of gram-negative bacteria. tonB mutants are impaired in virulence in various animal infection systems [44, 45]. However, some bacteria contain up to three tonB and exbB, exbD genes, which mjght participate in different iron uptake systems (see, for example, Iron Transport of Vibrio cholerae Related to Virulence). In addition, it is usually not known which iron transport system is important for proliferation at a specific infection site. Moreover, the iron limitation usually encountered in the human body could serve as an environmental signal that tells a bacterial strain its location in the human body. This could induce expression of genes required for multiplication, but might not be directly related to the iron supply. Therefore, different approaches are required to elucidate a relationship between iron transport and virulence. Such studies have involved knocking out a particular iron transport system and a genomewide search for the expression of genes in vivo compared to the expression of genes in synthetic media under iron-deplete and iron-replete conditions. Such large-scale expression profiles usually reveal genes related to the iron supply. These genes encode proteins for siderophore biosynthesis and transport, heme transport, hemolysins, and toxins. The most prominent toxin is the diphtheria toxin, which is synthesized under iron-limiting conditions. Other iron-regulated toxjns are the Shiga toxin of Shigella and E. coli strains, the hemolysins/ cytolysins of Serratia marcescens and certain E. coli strains, exotoxin A of Pseudomonas aeruginosa, and the tetanus toxin of Clostridium tetani. By damaging cells, the toxins can mobilize intracellular iron sources and make them available to bacteria. S. marcescens, for example, colonizes the intestine of Caenorhabditis elegans and kills the nematode. S. marcescens mutants are

Iron Transport Related to Virulence

215

impaired in virulence when they carry a transposon in the hemolysin gene or in a siderophore biosynthesis gene [46]. Stress by Iron Surplus Not only iron shortage, but also iron surplus can affect the outcome of a bacterial infection. Aerobic metabolism constantly creates hydrogen peroxide and superoxide radicals. If too much HzO z is formed, it might not be completely destroyed by catalase and peroxidase. In the Haber-Weiss reaction, the oxygen radical reacts with HzO z to form the highly reactive hydroxyl radical and hydroxyl anion. In the Fenton reaction, Fe z+ converts HzO z to the hydroxyl radical and hydroxide anion. Fe3+ oxidizes the oxygen radical to oxygen. HzO z, the oxygen radicals, and the hydroxyl radicals damage DNA, lipids in membranes, and proteins. The lack of regulation of iron metabolism could, therefore, be deleterious to cells [47]. This has been demonstrated for E. coli, in which a mutation in the fur (iron uptake regulator) gene renders cells sensitive to oxygen. An additional mutation in the recA gene, which is involved in DNA repair, kills cells when they are cultivated under oxic conditions [48]. The surplus of reactive intracellular free iron might result from an uncontrolled import and the lack of intracellular iron storage proteins. Iron uptake is controlled by the fur gene in most gram-negative bacteria and certain gram-positive bacteria with a low GC content and by the dtxR gene in most (GC-rich) gram-positive bacteria. When the intracellular iron concentration reaches a certain level, the Fur and DtxR proteins are loaded with Fe2+ and repress transcription of genes encoding iron transport proteins and enzymes that synthesize siderophores [7]. Two types of iron storage proteins contribute to intracellular iron homeostasis in bacteria [22]. Ferritins are also found in eukaryotes, and heme-containing bacterioferritins are only found in bacteria. Both types are composed of 24 identical subunits that form an almost spherical shell into which more than 2,000 FeH ions can be deposited. The FtnA ferritin of E. coli accumulates iron in the postexponential growth phase in the presence of excess iron in the medium and supports subsequent growth under iron-deficient conditions. Helicobacter pylori and Campy/abaeter jejuni express a siInilar protein that stores iron and protects cells against oxygen damage. No physiological role has been ascribed to the Bfr bacterioferritin of E. coli, but a bfr mutant of P aeruginosa is sensitive to peroxides. Dps is another iron-binding protein that forms a shell, but with 12 subunits. Dps is probably less important for iron storage than for protecting DNA against the combined action of iron and HzOz. Iron Transport orE. coli and Shigella Related to Virulence Pathogenic E. coli strains express ten outer membrane proteins that transport ferric siderophores and heme (table I). All the ferric hydroxamates (aerobactin,

Braun

216

Table 1. Iron transport systems of E. coli Substrate

Outer membrane protein

Periplasmic protein

Cytoplasmic membrane proteins

Enterobactin Salmochelin Catecholates Catecholates Ferrichrome Aerobactin Coprogen Citrate Heme Yersiniabactin d Fe 2 +

FepA IroN Cir Fiu FhuA IutA FhuE FecA ChuA FyuA

FepB FepB FepB FepB FhuD FhuD FhuD FecB ChuTe NI

FepD", FepG", FepC b FepD", FepG", FepC b FepD", FepG", FepC b FepDa, FepGa, FepC b FhuB a, FhuC b FhuB a, FhuC b FhuB", FhuC b FecC", FecD", FecE b ChuUa,C, Chuyb,c Ybtp' YbtP FeoB

aTransmembrane transport proteins in the cytoplasmic membrane. bATPase. CDesignations adapted from S. dysenteriae which is justified by the highly homologous E. coli and Shigella genomes. In E. coli K-12 ChuA alone is sufficient to support heme-dependent growth but the transport system in the cytoplasmic membrane may increase sensitivity to heme and rate of heme uptake. dThe transport system of yersiniabactin is encoded on pathogenicity islands which occur in various Enterobacteriaceae. The nomenclature of reference 58 was used. For further details, see text and references 8 and 49-51. NI = Not identified.

ferrichrome, coprogen) for which specific transporters are found in the outer membrane are transported by the same transport system across the cytoplasmic membrane. The same holds true for the ferric catecholates, including ferric enterobactin and presumably ferric salmochelin, which are transported across the cytoplasmic membrane by the same system. It is not clear whether or to what extent the entire FepBCD transport system is involved in the fenic salmocheJin transport. The heme transport system has been characterized in Shigella dysenteriae and its phylogenetic distribution in enteric bacteria has been determined [52]. The assignment of the heme genes to functions is based on the first functionally characterized heme transport system of Yersinia enterocolitica [53]. Heme and aerobactin transport, as well as TonE are required for virulence of the uropathogenic E. coli strain CFT073 in a mouse model of urinary tract infection [54]. In addition, E. coli strains isolated from patients with an intra-abdominal infection have been shown to secrete a protease, Hbp, that degrades hemoglobin. Hbp binds the released heme [55] and promotes the growth of Bacteroides ji'agilis, which is frequently

Iron Transport Related to Virulence

217

associated with E. coli in intra-abdominal infections. In a mouse infection model, Hbp contributes to the pathogenic synergy of these two organisms in abscess development. Heme transport systems are widely distributed among gram-positive and gram-negative bacteria [10, 11]. The Fe3+ -yersiniabactin transport system is frequently encoded on a 'high pathogenicity island', which occurs in several Enterobacteriaceae [56], but is also present in strains with less pathogenic potential [57]. The transport system of Fe3+ -yersiniabactin across the cytoplasmic membrane is interesting since the two permease proteins YbtP and YbtQ are each fused with the ATPase [58], as is found with human ABC export proteins. Subcutaneous infection by a ybtP mutant fails to cause disease in mice, a route that mimics Yersinia pestis transmission by fleas causing bubonic plague. To date there has been no association reported between virulence and the ferric citrate transport system, in which FecB (binding protein), FecCD (permease), and FecE (ATPase) catalyze transport across the cytoplasmic membrane. A nearly identical transport system is located on a pathogenicity island of Shigella flexneri [59]. Coliform isolates of E. coli and Klebsiella pneumoniae from bovine inflammatory infections (mastitis) contain FecA, as evidenced by anti-FecA antibodies [60], and FecA is being considered as a vaccine component for the treatment of mastitis. A study of the regulation of the ferric citrate transport proteins tillcovered a new type of transcription regulation. The inducer of the transcription of the transport genes binds to the FecA outer membrane protein and elicits a signal that is transmitted by FecA across the outer membrane to the FecR protein, which transmits the signal across the cytoplasmic membrane. In the cytoplasm, the FecI sigma factor is activated and directs the RNA polymerase specifically to the promoter ofthejec transport genes upstream ofjecA [61,62]. Siderophores like ferrichrome and coprogen, which are not synthesized by E. coli or any other bacteria, but which are transported by many bacteria, including E. coli, might be used during coinfection with fungi that synthesize the siderophores or during bacterial growth outside the human body. The large variety of transport systems for ferric siderophores and heme found in E. coli and Shigella are typical for pathogenic bacteria. The systems are dist:l;buted among bacteria by horizontal gene transfer. For example, the aerobactin synthesis genes are found on plasmids in E. coli and Salmonella, on pathogenicity islands in S. flexneri and Shigella sonnei, and on the chromosome of Shigella boydii and certain E. coli strains [8, 63]. Another example is the recently discovered iroN gene, which was originally identified in Salmonella enterica and then shown to contribute to the uropathogenicity of E. coli isolates [64, 65]. iroN is encoded on a pathogenicity island on the chromosome [64] and on a transmissible plasmid [65]. In a mouse model of ascending urinary tract infection, IroN contributes to colonization of the bladder, kidneys, and urine [64].

Braun

218

In addition to the Fe H transport systems, E. coli also contains an Fe H transport system, which is encoded by theftoAB genes [23]. This transport system functions under anoxic conditions, as found in the colon and in biofilms. Iron Transport of Salmonella Related to Virulence S. enterica serovar Typhimurium has iron transport systems similar to those of E. coli and Shigella, but so far no heme or ferric citrate transport system has been described. However, a heme transport gene operon similar to that in S. dysenteriae is encoded on the Salmonella typhimurium genome. The known systems include those related to the outer membrane transporters FhuA, FepA, FoxA, Cir, and IroN. An additional transport system presumably transports iron, as was first demonstrated for the sfuABC iron transport system of S. marcescens [66] and then for the jbpABC system of N. gonorrhoeae, hjbpABC (hitABC) of H. injluenzae [18], and yfuABC of Y. pestis [67]. sitA encodes a putative periplasmic permease, sitB an ATPase, and sitCD a permease [68]. However, sitABC is not homologous to the SfiIABC-type transport systems, but is homologous to yfeABC of Y. pestis and it transports Mn2+ with a much higher affinity than Fe H . The Sit system is widely distributed in all S. enterica serovars and is required for full virulence of S. typhimurium [69]; the Yfe system is essential for virulence of Y. pestis [70]. Iron transport systems are redundant, depending on the test system, since depleting one system may have no effect on bacterial virulence. The S. enterica genome also carries theftoAB genes, which encode an FeH transport system. Single mutations of sitA, feoB, or iucD (Fe H -aerobactin transport) in S. jlexneri do not impair the growth of these bacteria on a Henle cell monolayer; however, triple mutants do not form plaques [71]. A novel siderophore, designated salmochelin, was discovered only recently in S. enterica serovar Typhimurium LT2. The iroB gene product, encoded in the iron-regulated gene cluster iroNEDCB, shows sequence similarity to glycosyl transferases. This finding prompted a search for the function of lroB. Indeed, lroB was shown to encode an enzyme that glucosylates enterobactin at the 5' position of the benzoyl ring, forming a C-C bond [106]. The published tentative structure carries the two glucosyl moieties inserted between two 2,3-dihydroxybenzoylserine residues [49]. In a Salmonella culture, salmochelin is more abundant and is more soluble than enterobactin. Therefore, it might be less able to elicit antibodies than enterobactin, which serves, bound to serum albumin, as a hapten. Transport of Fe3 +-salmochelin across the outer membrane is mediated by IroN and to a lesser extent by the FepA and Cir transporters. Iron Transport ofP. aeruginosa Related to Virulence Pyoverdin and pyochelin are two well-studied siderophores that supply iron to P. aeruginosa. A number of indications show a relationship between

Iron Transport Related to Virulence

219

iron supply and virulence of P aeruginosa in animal infection models: derepression of siderophore synthesis genes, synthesis of the siderophores pyoverdin and pyochelin and the related transport proteins, release of iron from the host iron-binding proteins transferrin and lactoferrin, and reduction of virulence of mutants deficient in synthesis of siderophores or Fe H siderophore transport proteins. In addition, exotoxin A synthesis is controlled by the iron supply via the Fur repressor. A tonB mutant devoid of Fe 3 + uptake via pyoverdin, pyochelin, and heme grows in the muscles and lungs of immunosuppressed mice, but does not kill the animals [72]. Pyoverdin- and pyochelin-negative double mutants multiply, but do not kill the mice; however, intranasal inoculation of wild-type bacteria results in multiplication and killing [73]. PvdS (see below) is an ECF sigma factor synthesized in chronic lung infections affiliated with cystic fibrosis and contributes to the synthesis of exotoxin A [74]. Complex regulatory devices underlie iron-mediated control of gene expression in P aeruginosa. For example, iron-loaded Fur does not bind to the promoter of the toxA gene of exotoxin A, but acts via the pvdS gene product, which regulates 26 iron-repressible genes. pvdS encodes an ECF sigma factor of the FecI type (see Iron Transport of E. coli and Shigella Related to Virulence), and its synthesis is repressed by binding of Fe 2 + -Fur to the pvdS promoter [75]. The activity of PvdS is controlled by pyoverdin secreted in the growth medium; pyoverdin (probably Fe H -pyoverdin) binds to the FpvA protein in the outer membrane. FpvA displays several functions: it acts as a signal receiver and as a signal transmjtter across the outer membrane, and it transports FeH -pyoverdin across the outer membrane. The signal is transmitted by the FpvR protein across the cytoplasmic membrane into the cytoplasm, where PvdS is converted into an active sigma factor. Since PvdS is active in mutants lacking FpvR and overexpression of FpvR inactivates PvdS, FpvR probably functions as an anti-sigma factor of PvdS [75]. PvdS directs the RNA polymerase to the promoter of the iron-repressible genes, including the pyoverdin synthesis genes. jpvR transcription is repressed by Fe 2+ -Fur, as is transcription of a second ECF sigma factor gene,fpvI. FpvI synthesis is regulated like PvdS synthesis via Fe H -pyoverdin, FpvA, and FpvR, and controls synthesis ofFpvA. Heme uptake by P aeruginosa is mediated by two systems, one of which is encoded by the phuRSTUVW genes (fig. 2) [76]. This system is very similar to the heme transport system of Y enterocolitica. Heme is bound to the PfuR outer membrane protein that transports heme across the outer membrane. Further transport into the cytoplasm is achieved by an ABC transporter. The other heme transport system is similar to the heme transport system of S. marcescens and involves a hemophore that is secreted, releases heme from hemoglobin, and delivers it to the outer membrane transport protein (fig. 2).

Braun

220

hemP

R



Yersinia enterocolitica

Y P

hmuX

」エ^ Z セ

Yersinia pestis

R

phuR



Pseudomonas aeruginosa shuS

I



I

x

w

T

A

セi

Shigella dysenteriae

I

Y

V

U

II

hulA Vibrio cho!erae

tonBI exbBI hutB

0

C

ClIc:J:::1セ i N セ exbDI hxuC

hxuB

c::::J::1::J:I:::J:IQセjZ

Haemophilus influenzae

hemO

has!

hmbR

I

I

rhu!

R

I

bhuR

".

I

E

0

I •

I

B

E

STU

F

-l>II- -1B

F

⦅セi ⦅

Mセ

V

C'

hasR

(ECF')

Pseudomonas aeruginosa

I

I

0

A

I

hpuB I

I I.e:t>-- _ ...

I

Q>c:::::t> ::J:::::J:I=:II=:ICJII:•• セ 」 Z Z サ ^

Bordetella pertussis

hpuA

c::::::::::(>i

I 1.Q>ww _ . . .

I

hasR

I

/I A

R

S

Q:>q>!

Yersinia pestis

hgpA I

1 ::J:I::::J:::IiJ. .D..... t>I:::JIc:::J::

Neisseria meningitidis

Serratia marcescens

hxuA

i O^ャェZ Z ] セ ェZ]M⦅j

ct>c:::t>'

I

I

A

c:::{>

".

0

QN・Zエ^Mcャ⦅BGセji

I

....

E

...=-....t F

c:::{>

Outer membrane receptor; periplasmic binding protein; integral membrane protein; ATPase subunit;

c::::[>

-

-

..



Hemophore; hemophore secretion system; protein for heme utilization

Fig. 2. Heme transport systems of gram-negative bacteria. The upper panel shows the transport genes and some promoters (P). In the lower panel, genes for hemophore synthesis, secretion, and regulation, and not the actual heme transport genes are shown for S. marcescens, Y. pestis, and P aernginosa. The HasA hemophores are secreted by the type I secretion mechanism catalyzed by the proteins HasD, HasE, and HasF. HasB is structurally and functionally a TonE-like protein. has! and hasS, and rhu! and rhuR encode a transcription-signaling device of the FeclR type in which the I proteins represent eXlTacytoplasmic membrane (ECF) sigma factors that receive signals from outside the cytoplasm and the R or S protein transfers the signals across the cytoplasmic membrane. In Borde/ella pertussis, rhulR regulates transcription of the bhuRSTUV heme transport genes [for further information, see 10, 62, 77].

Iron Transport Related to Virulence

221

In S. marcescens, regulation of heme transport gene transcription is mediated by a signaling device of the FecIRA type [77]. Heme-loaded hemophore binds to the HasR heme transporter and induces transcription of the hasR gene via HasI, which functions as an ECF sigma factor, and HasS, which acts as an anti-sigma factor. Since P aeruginosa contains genes homologous to those in S. marcescens and arranged similarly, it is likely that the two Has regulatory systems function similarly. Analysis of the genome of P aenlginosa predicts nine additional regulatory devices of the FeclRA, HasISR, FpvA/FpvI, FpvR, and PvdS type. These systems usually have the same gene arrangement as ftcJRA, and the outer membrane proteins contain an extended amino-terminus, which in FecA interacts with FecR [12-14]. In addition to smface signaling elicited by the iron substrates, P aenlginosa controls iron usage by a number of additional regulatory mechanisms. For example, pyochelin synthesis and uptake is repressed by Fe2 + -Fur, which binds to promoters of the synthesis and uptake genes. The regulatory protein PchR acts as a repressor in the absence of pyochelin and as an activator in the presence of pyochelin [78]. Regulation of ferric enterobactin usage is mediated by a two-component system consisting of the PfeS signal receiver and the PfeR response regulator. Ferric enterobactin in the periplasm binds to the PfeS sensor kinase, which is autophosphorylated and transfers the phosphate group to the receiver domain of PfeR. Phosphorylated PfeR functions as a transcription activator of the pfeA gene, which encodes the high-affinity PfeA outer membrane transporter [79]. In this iron transport system and in all the other iron transport systems studied in P aeruginosa, the transported substrate induces synthesis of the cognate transport system. This is achieved by various mechanisms, but always results in the economic adaptation of the cells to the available iron somce. If only iron depletion of the Fm protein would derepress gene transcription, many of the approximately 13 iron transport systems would be synthesized, even though only the one for the available iron somce would be required.

Iron Transport of Vibrio cholerae Related to Virulence Three heme transport systems have been identified in V cholerae, represented by the outer membrane transporters HutA, HutR, and HasR [80]. A hutA hutR double mutant is impaired, but not completely unable to use hemin as an iron source. The triple mutant hutA hutR hasR is completely devoid of heme utilization. V cholerae HasR is similar to the HasR proteins of S. marcescens and P aeruginosa, which receive heme from the hemophore that releases heme from hemoglobin. In addition to the use of heme via transporters across the outer and cytoplasmic membranes, V cholerae can use the iron complexes of

Braun

222

the siderophores vibriobactin, enterobactin, and ferrichrome [81]. The transporters are preferentially coupled to one of the two TonB proteins present in V cholerae [82]. HasR, VctA, and IrgA, the latter two transport Fe3+ -enterobactin [83], are only coupled to TonB2, whereas HutA, HutR, ViuA (Fe3+ -vibriobactin transporter) and FhuA (ferrichrome transporter) can use TonB I and TonB2 [80]. In an infant mouse model, the triple mutant competes with the wild-type strain, which indicated additional iron sources in vivo [80]. Analysis of gene transcription in the rabbit ileal loop model have revealed enhanced transcription of heme and Fe 3 + transport genes and of the JeoAB genes, which encode an Fe2+transport system [84] that may have supplied the necessary iron.

Functions ojIron in Neisseria Related to Virulence A tonB mutant of Neisseria meningitidis does not actively transport iron and is unable to replicate within epithelial cells [85]. N gonorrhoeae and N meningitidis transport iron across the cytoplasmic membrane by an ABC transporter encoded by thejbpABC genes [18], which are similar to the sJuABC genes of S. marcescens, the hjbpABC (hi/ABC) of H. influenzae, and the yJABC genes of Y. pestis (see Iron Transport of Salmonella Related to Virulence). No siderophore seems to be involved in iron transport. In N gonorrhoeae and H. injluenzae, the iron might be delivered by the host transferrin and lactoferrin, which bind to highly specific outer membrane receptor proteins composed oftwo polypeptides: TbpA and TbpB for the transferrin receptor, and LbpA and LbpB for the lactoferrin receptor. The B components are lipoproteins and discriminate between iron-loaded and iron-unloaded transferrins and lactoferrins. The A components are similar to TonB-coupled ferric siderophore and heme transporters. TonB is not only required for the transport of iron across the outer membrane, but also for the release of Fe3+ from transferrin and lactoferrin [21]. The A and B components act in concert and interact with each other. Proteolytic degradation of TbpB is strongly influenced by coupling of TbpA to TonB. N gonorrhoeae mutants that lack the transferrin receptor do not elicit symptoms of urethritis in human male volunteers [86]. Two hemoglobin receptors have been identified in N meningitidis: a twocomponent receptor designated HpuAB and a one-component receptor designated HpmR. No siderophores have been identified in Neisseria. However, Neisseria can utilize Fe3+-enterobactin taken up via a TonB-coupled transporter across the outer membrane and an ABC transporter across the cytoplasmic membrane [5, 87]. Iron Transport oJStaphylococcus aureus Related to Virulence In S. aureus, several iron transport systems seem to operate. Ferrichrome is actively transported [88], and recently heme transport has been correlated with proteins (Isd) on the cell surface that are anchored to the murein by two sortases

Iron Transport Related to Virulence

223

[89]. S. aureus binds transferrin [90] and haptoglobin-hemoglobin [91]. In certain strains, slime production is enhanced by iron limitation [92]. Iron homoeostasis is regulated by the Fur repressor, whose synthesis is repressed by a homologous protein, PerR, which also regulates synthesis of the iron storage proteins ferritin and MrgA, a Dps homolog. PerR is required for full virulence of S. aureus in a murine skin abscess model [93]. The cell wall of S. aureus and Staphylococcus epidermidis contains the Tpn transferrin-binding protein, which is synthesized under iron-limiting growth conditions and elicits antibody fonnation in human serum and peritoneum upon staphylococcal infections [94]. The Tpn protein is the cell wall glyceraldehyde-3-phosphate dehydrogenase, which also binds plasmin [95]. It is assumed that the released iron is taken up into the cytoplasm by ABC transporters. Two such ABC transporters, encoded by the sirABC and sstABCD genes, have been partially characterized [96].

Fe3 + -Siderophores as Antibiotic Carriers Multidrug resistance against currently used antibiotics fonns an increasing problem in the treatment of bacterial diseases. One way out of the resistance dilemma is the development of new antibiotics. Since most antibiotics have been discovered during the decades of large-scale random screening, new strategies will have to be exploited. One possibility is the use of transport systems to transport antibiotics into cells. There are examples in which active transport, as opposed to diffusion, decreases the minimal inhibitory concentration (MIC) of an antibiotic more than 100-fold [97]. Antibiotics with Fe3+ -Hydroxamate Carriers Most antibiotics diffuse into bacteria, and their rate of diffusion and their activity at the target sites determine their efficiency, as measured by the MIe. In gram-negative bacteria, the outer membrane fonns an additional permeability barrier in addition to the cytoplasmic membrane, and renders gram-negative bacteria less sensitive to many antibiotics than gram-positive bacteria. However,1 if antibiotics are actively transported across the outer membrane, their MIC could be lower in gram-negative than in gram-positive bacteria because the antibiotics are accumulated in the periplasm and fonn a steep concentration graI dient into the cytoplasm, thereby enhancing the diffusion rate, or the antibiotio might even be actively transported across the cytoplasmic membrane. There are naturally occurring antibiotics that consist of an antibiotically active moiety and a siderophore carrier. The best-studied example is albomycin, which is composed of a trihydroxamate that binds Fe3+, a peptide linker, and a thioribosyl pyrimidine moiety that inhibits tRNA Ser synthetase [98]. Albomycin is

Braun

224

highly active toward gram-positive and gram-negative bacteria. The MlC against an E. coli strain is 200 times lower (0.05 f-Lg/ml) than of ampicillin (12.5 f-Lg/ml). The high specific activity comes from the active transport across the outer membrane and the cytoplasmic membrane into bacteria via the transport system of the structural analogue ferrichrome. The ferrichrome analogue serves as carrier of the antibiotically active thioribosyl pyrimidine group. After transport into the cytoplasm, iron is released from albomycin, and the thioribosyl pyrimidine group has to be cleaved from the carrier to be inhibitory. In E. coli, this is mainly achieved by peptidase N [1, 3, 97]. Mutants devoid of peptidase N activity are resistant to albomycin, and albomycin then serves as an iron carrier. Most of the thioribosyl pyrimidine moiety remains inside the cell, whereas the carrier is released into the culture medium. Albomycin is one of the very few antibiotics for which transport, intracellular activation, and target have all been characterized. Albomycin has been cocrystallized with FhuA to determine whether it binds to the ferrichrome binding site of FhuA and where the bulky side chain is located in FhuA (fig. 1). The crystal structure reveals that the Fe H -hydroxamate portion of albomycin occupies the same site on FhuA and is bound by the same amino acid side chains as ferrichrome [99]. The thioribosyl pyrimidine moiety binds in the external pocket via five residues that are not involved in ferrichrome binding. The crystal struchlre also reveals the hitherto unknown conformation of albomycin and the conformation in the transport-competent form. Unexpectedly, albomycin assumes two conformations in the crystal - an extended and a compact conformation. Both conformations fit into the external cavity of FhuA and occupy seven different amino acid ligands. The solvent-exposed external cavity of FhuA is sufficiently large to accommodate the voluminous side chain of albomycin. After transport across the outer membrane by FhuA, albomycin binds to FhuD in the periplasm. FhuD subsequently delivers albomycin to the permease in the cytoplasmic membrane. Cocrystals ofFhuD with bound albomycin have been obtained in sufficient quality to determine the structure [100]. In contrast to FhuA, where albomycin sits inside the molecule, in FhuD albomycin is exposed to the surface of the protein. The thioribosyl moiety is not even seen in the crystal since it is not fixed to the protein and is thereby flexible. The fixation of albomycin at the surface of FhuD explains the broader substrate specificity of FhuD in contrast to FhuA since space is less restricted at the protein surface than within a protein. Results of studies with albomycin demonstrate that the proteins involved in transport across the outer membrane and the cytoplasmic membrane tolerate substantial modifications of the substrate. The modular design of albomycin can be synthetically mimicked. Antibiotics that are ineffective because of poor entry

Iron Transport Related to Virulence

225

into the cells can be chemically linked to ferrichrome and then transported into cells as ferrichrome derivatives. CGP 4832 is a semisynthetic rifamycin derivative with an activity against many gram-negative bacteria 200-fold higher than that of unmodified rifamycin [101]. The reason for the increased activity of CGP 4832 is its energy-coupled transport by FhuA across the outer membrane of E. coli [10 I]. The use of FhuA as transporter is surprising since CGP 4832 does not contain iron and has no structural resemblance to ferrichrome or any other hydroxamate. To obtain insights into how CGP 4832 is transported by FhuA, the crystal structure of FhuA loaded with CGP 4832 was determined [102]. CGP 4832 occupies in FhuA largely the same site as ferrichrome (fig. 1). Nine residues that bind CGP 4832 also bind ferrichrome. Of 16 amjno acid residues that bind CGP 4832, 5 residues recognize those side chains of CGP 4832 in which it differs from unmodified rifamycin. Two additional amino acid residues specifically bind the unique CGP 4832 side chains, whereas the other residues bind to sites that CGP 4832 shares with rifamycin. The crystal structure reveals the conformation of CGP 4832, wmch demonstrates a completely different structure than that offerrichrome. Unlike albomycin, CGP 4832 is not transported via FhuBCD across the cytoplasmic membrane [101]. Rather, its active transport across the outer membrane results in an elevated concentration in the peri plasm, which facilitates diffusion across the cytoplasmjc membrane. It is the active transport across the outer membrane that reduces the MIC 200-fold. Salmycins have been isolated from Streptomyces violaceus 37290 (DSM 8286) and are highly active against staphylococci and streptococci (MIC 10 jJ.g/ml). Salmycins consist of an Fe3+ -siderophore with a ferrioxamine group and an antibiotically active aminodisaccharide, which in salmycin B consists of a 2-ketoglucose linked to the 2-position of a 6-methylamjnoheptopyranose [103]. It is assumed that the aminodisaccharide is released from the carrier by cleavage of the ester bond. Salmycins seem to inhibit protein synthesis by a yet unknown mechanism. Ferrimycins are among the first sideromycins discovered [97]. The action of ferrimycins is antagonized by ferroxamine B, which competes for ferrimycin uptake. Ferrimycin inmbits incorporation of amino acids into proteins of S. aureus SG511. Ferrimycin is difficult to isolate and for this reason has recently been studied less than albomycin and salmycin. Antibiotics with Fe3+ -Catecholate Carriers Enterobactin is the most prominent catecholate siderophore with an extremely high Fe 3 + stability constant. It consists of three dihydroxy benzoyl serine residues linked to a cyclic trimer by ester bonds. No natural Fe3+catecholates with antibiotic activity are known. However, chemically synthesized

Braun

226

catechol-substituted cephalosporins display MIC values below I !-1g/ml [104, 105], particularly against gram-negative bacteria, including P aeruginosa. Their antimicrobial activities can exceed the activity of the unsubstituted cephalosporins more than I DO-fold. Their high activity is related to their active transport into the periplasm, where the target, the murein transpeptidase, is located. They are transported across the outer membrane by the Fe H -catecholate transport proteins Fiu and Cir [26]. Iron limitation increases the susceptibility of E. coli strains since low iron derepresses Fiu and Cir synthesis.

Resistance to Fe H -Siderophore Antibiotics Resistant bacteria emerge on every nutrient agar plate containing antibiotics that are carried into the bacteria by active Fe H -siderophore transport systems. The higher the number of genes involved in a particular tmnsport system, the higher the frequency of resistance. However, when two transport systems are used by an antibiotic, for example Cir and Fiu for the cephalosporin catecholates, the frequency of resistant mutants is low. Although the high resistance frequency seems to prevent development of such antibiotics as antibacterial drugs, the in vivo situation might be quite different. In cases where an iron transport system is important for the proliferation of the pathogenic bacteria, loss of the iron transport system is detrimental. Even when several iron transport systems exist and only one is inactivated by resistance to a particular antibiotic, the inactivated system might be the one that is essential for the bacteria to survive and multiply at the site of infection in the human host. Under these circumstances, it does not matter whether the number of bacteria is reduced by the antibiotic or by loss of the iron supply since under both conditions the immune defense system gains time to cope with the infection.

Concluding Remarks

Iron deficiency was also designated nutritional immunity which meant that growth inhibition by lack of iron prevents bacterial multiplication. Lack of growth or growth retardation gives the natural and the adaptive immunity system the chance to cope with an infection. Iron is the only nutrient for which an essential role in growth of many bacterial pathogens causing various diseases in humans and animals has been demonstrated. There are certainly many more nutrients which playa decisive role in extra- and intracellular multiplication of bacteria. However, it is difficult to identify these nutrients. Large-scale expression profiles of metabolic genes in bacteria isolated from human patients without further culturing and from animal models may indicate metabolic pathways from which the nutrients may be derived. From a purely scientific point of view

Iron Transport Related to Virulence

227

the iron supply systems are of great interest with regard to the various ways insoluble Fe H is complexed by siderophores, heme, transferrin, and lactoferrin and transported into the bacterial cells by distinct and very sophisticated mechanisms. For the avoidance of iron shortage and iron surplus the transport systems are regulated by various means, iron-dependent repression, downregulation by small RNAs, transcription enhancement by two-component systems, and transcription initiation by surface signaling. In the future, a detailed knowledge of iron uptake and intracellular iron metabolism may be applied to interfere with bacterial growth as a means to control bacterial diseases, and siderophore antibiotics (sideromycins) may be used when treatment with other antibiotics fails because of resistance.

Acknowledgments I would like to thank Klaus Hantke for preparation of figure 2, Michael Braun for preparation of figure 1, and Karen A. Brune for critically reading the manuscript. The author's work was supported by the Deutsche Forschungsgemeinschaft (Forschergruppe 'Bakterielle Zeilliiille: Synthese, Funktion und Wirkort' , Br 330/14-2) and the Fonds der Chemischen Industrie.

References

2 3 4 5 6 7 8 9 10 II 12

Braun V, Hantke K, Koster W: Bacterial iron transport: Mechanisms, genetics, and regulation; in Sigel A, Sigel H (eds): Metal Ions in Biological Systems. New York, Marcel Dekker, 1998, vol 35, pp 67-145. Braun V: Iron uptake mechanisms and their regulation in pathogenic bacteria. Int J Med Microbiol 2001;291:67-79. Braun V, Braun M: Active transport of iron and siderophore antibiotics. Curr Opin Microbiol 2002;5: 194-20 I. Braun V: Iron uptake by Escherichia coli. Front Biosci 2003; 1:409-421. Rhode KH, Dyer DW: Mechanisms of iron acquisition by the human pathogens Neisseria meningitidis and Neisseria gonorrhoeae. Front Biosci 2003;8: 1186-1218. Bullen JJ, Griffith E: Iron and Infection. Molecular, Physiological and Clinical Aspects. New York, Wiley, 1999. Hantke K, Braun V: The art of keeping low and high iron concentrations in balance; in Storz G, Hengge-Aronis R (eds): Bacterial Stress Responses. Washington, ASM Press, 2000, pp 275-288. Crosa JH: The relationship of plasmid-mediated iron transport and bacterial virulence. Annu Rev Microbiol 1984;38:69-89. Crosa JH: Signal transduction and transcriptional and posttranscriptional control of iron-regulated genes in bacteria. Microbiol Mol Bioi Rev 1997;61:319-336. Wandersman C, Stojiljkovic I: Bacterial heme sources: The role of heme, hemoprotein receptors and hemophores. Curr Opin MicrobioI2000;3:215-220. Genco CA, Dixon OW: Emerging strategies in microbial haem capture. Mol Microbiol2001 ;39:1-11. Vasil ML, Ochsner UA: The response of Pseudonwnas aeruginosa to iron: Genetics, biochemistry and virulence. Mol Microbiol 1999;34:399-413.

Braun

228

13 14

15 16 17 18 19 20 21 22 23 24 25 26

27

28 29 30

31 32 33

34 35

36

37

Visca p, Leono L, Wilson MJ, Lamont IL: Iron transport and regulation, cell signaling and genomics: Lessons from Escherichia coli and Pseudomonas aeruginosa. Mol Microbiol 2002;45: 1177-1190. Martinez-Bueno MA, Tobes R, Rey M, Ramos JL: Detection of multiple extracytoplasmic function (ECF) sigma factors in the genome of Pseudomonas putida KT2440 and their counterparts in Pseudomonas aeruginosa PAOI. Environ Microbiol 2002;4:842-855. Koster W: ABC transporter-mediated uptake of iron, siderophores, heme and vitamin B 12 . Res MicrobioI2001;152:292-301. Clarke TE, Tari LW, Vogel HJ: Structural biology of bacterial iron uptake systems. Cun Top Med Chern 2001;1:7-30. Ratledge C, Dover LG: Iron metabolism in pathogenic bacteria. Rev Microbiol 2000;54:881-94 \. Mietzner TA, Tencza SB, Adhikari P, Vaughan KG, Nowalk AJ: Fe(llI) periplasm-to-cytosol transporters of gram-negative pathogens. Curr Top Microbiol Immunol 1998;225: 113-135. Woolridge KG, Williams PH: Iron uptake mechanisms of pathogenic bacteria. FEMS Microbiol Rev 1993;12:325-348. Schryvers AB, Bonna R, Yu RH, Wong H, Retzner M: Bacterial lactoferrin receptors. Adv Exp Med BioI 1998;443: 123-133. Cornelissen CN: Transferrin-iron uptake by gram-negative bacteria. Front Biosci 2003;8:836-847. Andrews SC, Robinson AK, Rodriguez-Quinones F: Bacterial iron homoestasis. FEMS Microbiol Lett 2003;27:215-237. Hantke K: Is the bacterial ferrous iron transporter FeoB a living fossil? Trends Microbiol 2003;11:192-195. Perry RD, Abney J, Mier I Jr, Lee Y, Bearden SW, Fetherston JD: Regulation of the Yersinia pestis Yfe and Ybt iron transport systems. Adv Exp Med Bioi 2003;529:275-283. Davidson AL: Mechanism of coupling of transport to hydrolysis in bacterial ATP-binding cassette transporters. J Bacteriol 2002; 184: 1225-1233. Bruns C, Nowalk A, Arvai A, McTigue M, Vaughan K, Mietzner T, McRee D: Structure of Haemophilus injluenzae Fe+ 3 -binding protein reveals convergent evolution within a superfamily. Nat Stmct Bioi 1997;4:919-924. Clarke TE, Braun V, Winkelmann G, Tari LW, Vogel I-IJ: X-ray crystallographic structures of the Escherichia coli periplasmic protein FhuD bound to hydroxamate-type siderophores and the antibiotic albomycin. J BioI Chern 2002;277:13966-13972. Locher KP, Lee AT, Rees DC: The E. coli BtuCD structure: A framework for ABC transporter architecture and mechanism. Science 2002;296: I091-1 098. Borths EL, Locher KP, Lee AT, Rees DC: The structure of Escherichia coli BtuF and binding to its cognate binding cassette transporter. Proc Nat! Acad Sci USA 2002;99: 16642-16647. Groeger W, Koster W: Transmembrane topology of the two FhuB domains representing the hydrophobic components of bacterial ABC transporters involved in the uptake of siderophores, haem and vitamin B12. Microbiology 1998;144:2759-2769. Nikaido H: Prevention of drug access to bacterial targets: Permeability barriers and active efflux. Science 1994;264:382-388. Bradbeer C: The proton motive force drives the outer membrane transport of cobalamin in Escherichia coli. J BacterioI1993;175:3146-3150. Braun V: Energy-coupled transport and signal transduction through the gram-negative outer membrane via TonB-ExbB-ExbD-dependent receptor proteins. FEMS Microbiol Rev 1995;16: 295-307. Postle K, Kadner RJ: Touch and go: Tying TonB to transport. Mol Microbiol 2003;49:869-882. Ferguson AD, Hofmann E, Coulton JW, Diederichs K, Welte W: Stmctural basis for siderophoremediated iron transport: Crystal structure of FhuA with bound lipopolysaccharide. Science 1998;282:2215-2220. Locher KP, Rees B, Koebnik R, Mitschler A, Moulinier L, Rosenbusch Jp, Moras D: Transmembrane signaling across the ligand-gated FhuA receptor: Crystal structures of free and ferrichrome-bound states reveal allosteric changes. Cell 1998;95:771-778. Buchanan SK, Smith BS, Venkatramani L, Xia D, Esser L, Palnitkar M, Chakraborty R, van der Helm D, Deisenhofer J: Crystal structure of the outer membrane active transporter FepA from Escherichia coli. Nat Struct Bioi 1999;6:56-63.

Iron Transport Related to Vimlence

229

38 39

40 41

42 43 44 45

46

47 48

49

50

51 52

53 54

55

56

57

58

Ferguson AD, Chakraborty R, Smith BS, Esser L, van der Helm D, Deisenhofer J: Structural basis of gating by the outer membrane transporter FecA. Science 2002;295: 1658-1659. Wyatt W, Grizot S, Buchanan SK: Structural evidence for iron-free citrate and ferric citrate binding to the TonS-dependent outer membrane transporter FecA. J Mol Bioi 2003;332: 353-368. Merianos HJ, Cadieux N, Lin CH, Kadner RJ, Cafiso DS: Substrate-induced exposure of an energycoupling motif of a membrane transporter. Nat Struct Bioi 2000;7:205-209. Cadieux N, Kadner RJ: Site-directed disulfide bonding reveals an interaction site between energycoupling protein TonB and BtuB, the outer membrane cobalamin transporter. Proc Natl Acad Sci USA 1999;96: I0673-1 0678. Ogierman M, Braun V: Interactions between the outer membrane ferric citrate transporter FecA and TonS: Studies of the FecA TonB box. J Bacteriol 2003; 185: 1870-1885. Cimento Dp, Mohanty AK, Kadner RJ, Wiener MC: Substrate-induced transmembrane signaling in the cobalamin transporter BtuB. Nat Struct Bioi 2003; J0:394-401. Reeves SA, Torres AG, Payne SM: TonB is required for intracellular growth and virulence of Shigella dysenteriae. Infect Jrnmun 2000; II :6329--6336. Torres AG, Redford P, Welch RA, Payne SM: TonS-dependent systems of uropathogenic Escherichia coli: Aerobactin and heme transport and TonB are required for virulence in the mouse. Infect Irnmun 200 1;69:6179--6185. Kurz CL, Chavet S, Andres E, Aurouze M, Vallet J, Michel GP, Uh M, Celli J, Filloux A, de Bentzmann S, Steinmetz I, Hoffmann JA, Finlay BB, Gorvel J-P, Ferrandon D, Ewbank JJ: Virulence factors of the human opportunistic pathogen Serratia marcescens identi fied by in vivo screening. EMBO J 2003;22:1451-1460. Braun V: Avoidance of iron toxicity through regulation of bacterial iron transport. Bioi Chem 1997;378:779-786. Touati D, Jacques M, Tardat B, Bouchard L, Despied S: Lethal oxidative damage and mutagenesis are generated by iron in filr mutants of Escherichia coli: Protective role of superoxide dismutase. J Bacteriol 1995:177:2305-2314. I-Iantke K, Nicholson G, Rabsch W, Winkelmann G: Salmochelins, siderophores of Salmonella enterica and uropathogenic Escherichia coli strains, are recognized by the outer membrane receptor IroN. Proc Natl Acad Sci USA 2003;100:3677-3682. Schubert S, Rakin A, Karch H, Carniel E, Heesemann J: Prevalence of the 'high-pathogenicity island' of Yersinia species among Escherichia coli strains that are pathogenic to humans. Infect Irnmun 1998;66:480-485. Torres AG, Payne SM: Haem transport in enterohemorrhagic Escherichia coli 0157:H7. Mol Microbiol 1997;23:825-833. Wyckoff EE, Duncan D, Torres AG, Mills M, Maase K Payne SM: Structure of the Shigella dysenteriae haem transport locus and its phylogenetic distribution in enteric bacteria. Mol Microbiol 1998;28: I 139-1 152. Stojiljkovic I, Hantke K: Hemin uptake systems of Yersinia enterocolitica: Similarities with other TonB-dependent systems in gram-negative bacteria. EMBO J 1992; J 1:4359-4367. Torres AG, Redford P, Welch RA, Payne SM: TonB-dependent systems ofuropathogenic Escherichia coli: Aerobactin and heme transport and TonB are required for virulence in mice. Infect Immun 2001 ;69:6179--6185. Otto BR, van Dooren SJ, Dozois CM, Luirink J, Oudega B: Escherichia coli hemoglobin contributes to synergistic abscess formation and heme-dependent growth of Bacteroidesfragilis. Infect Immun 2002;70:5-10. Schubert S, Picard B, Gouriou S, Heesemann J, Denamur E: Yersinia high-pathogenicity island contributes to virulence in Escherichia coli causing extraintestinal infections. Infect Immun 2002;70:5335-5337. Oelschlager TA, Zhang D, Schubert S, Carniel E, Rabsch W, Karch H, Hacker J: The highpathogenicity island is absent in human pathogens of Salmonella enterica subspecies I but present in isolates of subspecies III and VI. J BacterioI2003;185:J007-1 J II. Fetherton JD, Bertolino VJ, Perry RD: YbtP and YbtQ: Two ABC transporters required for iron uptake in Yersinia pestis. Mol Microbiol 1999;32:289-299.

Braun

230

59

60

61 62 63 64

65

66

67 68 69

70 71

72 73 74

75

76 77

78

79 80

Luck SN, Turner SA, Rajakumar K, Sakellaris H, Adler B: Ferric dicitrate transport system (Fec) of Shigellaflexneri 2a YSH6000 is encoded on a novel pathogenicity island carrying multiple antibiotic resistance genes. Infect Immun 2001 ;69:601 2---{)02 I. Lin J, Hogan JS, Smith KL: Antigenic homology of the inducible ferric citrate receptor (FecA) of coliform bacteria isolated from herds with naturally occurring bovine intrarnammary infections. Clin Diagn Lab Immunol 1999;6:966-969. Braun V: Surface signaling: Novel transcription initiation mechanism starting from the cell surface. Arch Microbiol 1997;237:325-331. Braun V, Mahren S, Ogierman M: Regulation of the Feci-type ECF sigma factor by transmembrane signalling. CUff Opin Microbiol 2003;6: 173-180. Vokes SA, Reeves SA, Torres AG, Payne SM: The aerobactin iron transport system genes in Shigella .flexnei are present within a pathogenicity island. Mol Microbiol 1999;33:63-73. Russo TA, McFadden CD, Carlino-MacDonald UB, Beanan JM, Barnard n, Johnson JR: IroN functions as a siderophore receptor and is a urovirulence factor in an extraintestinal pathogenic isolate of Escherichia coli. Infect [mmun 2002;70:7156-7160. Sorsa LJ, Dufke S, Heesemann J, Schubert S: Characterization of an iroBCDEN gene cluster on a transmissible plasmid of uropathogenic Escherichia coli: Evidence for horizontal transfer of a chromosomal virulence factor. Infect Immun 2003;71 :3285-3293. Angerer A, Gaisser S, Braun V: Nucleotide sequences of the sfitA, sfitB, and sfitC genes of Serratia marcescens suggest a periplasmic-binding-protein-dependent iron transport mechanism. J Bacteriol 1990; 172:572-578. Gong S, Bearden Sw, Geoffroy VA, Fetherston JD, Perry RD: Characterization of the Yersinia pestis Yfu ABC inorganic iron transport system. Infect Immun 200 1;67:2829-2837. Zhou 0, Hardt WD, Galan JE: Salmonella typhimurium encodes a putative iron transport system within the centisome 63 pathogenicity island. Infect Immun 1999;67: 1974-1981. Janakiraman A, Siauch JM: The putative iron transport system sitABCD encoded on SPII is required for full virulence of Salmonella typhimurium. Mol Microbiol 2000;35: 1146-1155. Bearden SW, Perry RD: The Yfe system of Yersinia pestis transports iron and manganese and is required for full virulence of plague. Mol Microbiol 1999;32:403--414. Runyen-Janecky LJ, Reeves SA, Gonzales EG, Payne SM: Contribution of the Shigella flexneri Sit, luc, and Feo iron acquisition systems to iron acquisition in vitro and in cultured cells. Infect lmmun 2003;71:1919-1928. Takase H, Nitanai H, Hoshino K, Otani T: Impact of siderophore production on Pseudomonas aentginosa infections in immunosuppressed mice. Infect lmmun 2000;68: 1834-1839. Takase H, Nitanai H, Hoshino K, Otani T: Requirement of the Pseudomonas aeruginosa tonB gene for high-affinity iron acquisition and infection. Infect Immun 2000;68:4498--4504. Hunt TA, Peng WT, Loubens I, Storey DG: The Pseudomonas aeruginosa alternative sigma factor PvdS controls exotoxin A expression and is expressed in lung infections associated with cystic fibrosis. Microbiology 2002; 148:3183-3193. Beare PA, For RJ, Martin LW, Lamont IL: Siderophore-mediated cell signaling in Pseudomonas aentginosa: Divergent pathways regulate virulence factor production and siderophore receptor synthesis. Mol MicrobioI2003;47:195-207. Ochsner UA, Johnson Z, Vasil ML: Genetics and regulation of two distinct haem uptake systems, pjIJ and has, in Pseudomonas aeruginosa. Microbiology 2000; 146: 185-198. Rossi MS, Paquelin A, Ghigo JM, Wandersman C: Haemophore-mediated signal transduction across the bacterial cell envelope in SelTatia marcescens: The inducer and the transported substrate are different molecules. Mol Microbiol2003;48:1467-1480. Heinrichs DE, Poole K: PchR, a regulator of ferric pyochelin receptor gene UptA) expression in Pseudomonas aeruginosa, functions both as an activator and as a repressor. J BacterioI1996;175: 5882-5889. Dean CR, Neshat S, Poole K: PfeR, an enterobactin-responsive activator of ferric enterobactin receptor gene expression in Pseudomonas aeruginosa. J Bacteriol 1996; 178:5361-5369. Mey AR, Payne SM: Haem utilization in Vibrio cholerae involves multiple TonS-dependent haem receptors. Mol Microbiol 200 1;42:835-849.

Iron Transport Related to Virulence

231

81 82

83

84

85 86

87 88 89

90 91

92

93

94

95

96 97 98

99

100

10 I

Rogers MB, Sexton JA, DeCastro GJ, Calderwood SB: Identification of an operon required for ferrichrome iron utilization in Vibrio cholerae. J Bacteriol 2000; 182:2350-2353. Occhino DA, Wyckoff EE, Henderson DP, Wrona TJ, Payne SM: Vibrio cholerae iron transport: Haem transport genes are linked to one of two sets of tonB, exbB, exbD genes. Mol Microbiol 1998;29: 1493-1507. Mey AR, Wyckoff EE, Oglesby AG, Rab E, Taylor RK, Payne SM: Identification of the Vibrio cholerae enterobactin receptors VctA and IrgA: IrgA is not required for virulence. Infect Immun 2002;70:3419-3426. XU Q, Dziejman M, Mekalanos JJ: Determination of the transcriptome of Vibrio cholerae during intraintestinal growth and midexponential phase in vitro. Proc Natl Acad Sci USA 2003; I00: 1286-1291. Larson JA, Higashi DL, Stojiljkovic I, So M: Replication of Neisseria menigitidis within epithelial cells requires TonB-dependent acquisition of host cell iron. Infect Irnmun 2002;70: 1461-1467. Cornelissen CN, Kelley M, Hobbs MM, Anderson]E, Cannon JG, Cohen MS, Sparling RF: The transferrin receptor expressed by gonococcal strain FA 1090 is required for the experimental infection of human male volunteers. Mol Microbiol 1998;27:611--616. Biegel Carson SD, Klebba PE, Newton SMC, Sparling PF: Ferric enterobactin binding and utilization by Neisseria gonorrhoeae. J Bacteriol 1999; 181 :2895-290 I. Sepulsky MT, Heinrichs DE: Identification and characterization ofjlwDJ andjhuD2, two genes involved in iron-hydroxamate uptake by Staphylococcus aureus. J Bacteriol 2001; 183:4994-5000. Mazmanian SK, Skaar Ep, Gaspar AH, Humayun M, Gornicki P, Jelenska J, Joachmiak A, Missiakas DM, Schneewind 0: Passage of heme-iron across the envelope of Staphylococcus aureus. Science 2003;299:906-909. Taylor JM, Heinrichs DE: Transferrin binding in Staphylococcus aureus: Involvement of a cell wall-anchored protein. Mol MicrobioI2002;43:1603-1614. Dryla A, Gelbmann D, von Gabain A, Nagy E: Identification of a novel iron regulated staphylococcal surface protein with haptoglobin-haemoglobin binding activity. Mol Microbiol2003; 49:37-53. Baldassarri L, Bertuccini L, Ammendolia MG, Arciola CR, Montana L: Effect of iron limitation on slime production by Staphylococcus aureus. Eur J Clin Microbiol Infect Dis 2001;20: 343-345. Horsburgh MJ, Clements MO, Crossley H, Ingham E, Foster SJ: PerR controls oxidative stress resistance and iron storage proteins and is required for virulence in Staphylococcus aureus. Infect Immun 2001 ;69:3744-3754. Modun BJ, Cockayne A, Finch R, Williams P: The Staphylococcus aureus and Staphylococcus epidermidis transferrin-binding proteins are expressed in vivo during infection. Microbiology 1998;144:1005-1012. Modun BJ, Evans RW, Joannou CL, Williams P: Receptor-mediated recognition and uptake of iron from human transferrin by Staphylococcus aureus and Staphylococcus epidermidis. Infect Immun 1998;66:3591-3596. Morrissey JA, Cockayne A, Hill PJ, Williams P: Molecular cloning and analysis of a putative siderophore ABC transporter from Staphylococcus aureus. Infect Immun 2000;68:6281-6288. Braun V: Active transport of siderophore-mimicking antibacterials across the outer membrane. Drug Resist Updat 1999;2:363-369. Stefanska AL, Fulston M, Houge-Frydrych CS, Jones JJ, Warr SR: A potent seryl tRNA synthetase inhibitor SB-217452 isolated from a Streptomyces species. J Antibiot 2000;53: 1346-1353. Ferguson AD, Braun V, Fiedler HP, Coulton JW, Diederichs K, Welte W: Crystal structure of the antibiotic albomycin in complex with the outer membrane transporter FhuA. Protein Sci 2000;9: 956-963. Clarke TE, Braun V, Winkelmann G, Tari LW, Vogel HJ: X-ray crystallographic structures of the Escherichia coli periplasmic protein FhuD bound to hydroxamate-type siderophore and the antibiotic aibomycin. J Bioi Chem 2002;277:13966-13972. Pugsley PA, Zimmerman W, Wehrli W: Highly efficient uptake ofa rifamycin derivative via FhuATonB-dependent uptake route in Escherichia coli. J Gen Microbiol 1987; 133:3505-3511.

Braun

232

102

103

104

105

106

Ferguson AD, Kodding J, Walker G, Bos C, Coulton JW, Diederichs K, Braun V, Welte W: Active transport of an antibiotic rifamycin derivative by the outer-membrane protein FhuA. Structure 2001;9:707-716. Vertesy L, Aretz W, Fehlhaber I-IW, Kogler 1-1, Salmycins AD: Antibiotics from Streptomyces violaceus DSM 8286 having a siderophore-aminoglycoside structure. Helv Chim Acta 1995;78: 46-60. Dolence EK, Minnick AA, Lin CE, Miller M: Synthesis and siderophore and antibacterial activity of NS-acetyl-N'-hydroxy-L-ornithine-derived siderophore-I3-lactam conjugates: lron-transport-mediated drug delivery. J Med Chern 1991 ;34:868-978. Curtiss NAC, Eisenstadt RL, East SJ, Cornford RJ, Walker LA, White AJ: Iron regulated outer membrane proteins of Escherichia coli K-12 and mechanisms of action of catechol-substituted cephalosporins. Antimicrob Agents Chemother 1988;32: 1879-1886. Bister B, Bischoff D, Nicholson GJ, Valdebenito M, Schneider K, Winkelmann G, Hantke K, Siissmuth RD: The structure of salmochelins: C-glucosylaled enterobaclins of Salmonella enterica. BioMetals 2004;17:471--481.

Prof. Volkmar Braun Mikrobiologie/Membranphysiologie, Universitiit Tiibingen Auf der Morgenstelle 28, DE-72076 Tiibingen (Germany) Tel. +497071 2972096, Fax +4970712975843 E-Mail [email protected]

Iron Transport Related to Virulence

233

Signaling and Gene Regulation Russell W, HelWald H (eds): Concepts in Bacterial Virulence. Contrib Microbial. Basel, Karger, 2005, vol 12, pp 234-254

Pathogenicity Islands and Their Role in Bacterial Virulence and Survival Bianca Hochhut, Ulrich Dobrindt, Jorg Hacker Institut fUr molekulare Infektionsbiologie, Universitat Wiirzburg, Wi.irzburg, Germany

Infections caused by bacterial pathogens are still a significant problem in modern medicine. Therefore, the identification of the factors that are related to the infections and the understanding of the processes involved in the evolution of pathogenic bacteria from their nonpathogenic progenitors is an important subject of research. It has long been known that acquisition of virulence determinants by horizontal gene transfer is one of the major driving forces in the emergence and evolution of new pathogens [reviewed in 1--4]. Furthermore, our knowledge of the organization of the bacterial genome has greatly increased within the last few years due to the availability of more than 120 completely sequenced eubacterial genomes, including those of almost all pathogenic bacteria, which has introduced a new area of pathogen research. It has become evident that the typical bacterial genome consists of a conserved 'core gene pool' comprising genes that encode essential structural features and fundamental metabolic pathways, and a 'flexible gene pool' that is more variable and encodes functions only advantageous under specific growth conditions. Core genes are characterized by a relatively homogenous G + C content and they are normally encoded in stable regions of the chromosome that are conserved in their organization in closely related species. In contrast, the flexible gene pool comprises variable regions of the chromosome and various mobile genetic elements such as plasmids, bacteriophages, IS elements and transposons, conjugative transposons, integrons and superintegrons that are transferred between different organisms by the means of natural transformation, transduction or conjugation. Many of the genes encoding toxins, adhesins, secretion systems, invasins or other virulence-associated factors have been found to be encoded by mobile genetic elements [overviews in 5, 6]. Furthermore, the analysis of the genomes

of closely related species has revealed that the conserved chromosomal backbone is interspersed with large regions that exhibit features of former mobile genetic elements that have been termed genomic islands (GEls) [7, 8]. GEls are broadly distributed and seem to be a common theme in most bacterial genomes. Originally, such elements were identified in uropathogenic Escherichia coli strains and were designated 'pathogenicity islands' (PAIs), because they encoded key virulence factors of these bacteria [9]. However, when regions with similar features were also found in nonpathogenic bacteria where they encoded other accessory functions, it was recognized that these elements are not limited to bacterial pathogens, but are present in most bacteria that have been analyzed. In this chapter, the role of GEls in bacterial virulence and survival will be discussed.

The Concept of GEls

Features of GEls A comparative analysis of microbial genome sequences has revealed that bacterial genomes can harbor variable and frequently significant amounts of foreign DNA [3]. The genome size of different variants of the same species or closely related species can vary by more than one megabase, which can be accounted for by the acquisition of large blocks of DNA such as plasmids, bacteriophages and GEls, as well as by the acquisition of smaller pieces of foreign DNA that have been described as 'islets'. Generally, GEls represent distinct pieces of DNA that have most of the following features in common suggesting that they originate from events of lateral gene transfer [10]. (I) GEls are present in the genomes of many bacteria but absent from the genomes of closely related strains or species. (2) GEls occupy relatively large regions of the chromosome and can cover between 10 and more than 100 kb, which may reflect the introduction of large pieces of DNA into a new host by horizontal gene transfer. Some strains also carry smaller pieces of DNA (1-10 kb) that have been termed 'genomic islets' in contrast to the larger islands. (3) GEls differ in their G + C content and their codon usage from that of the conserved regions of the chromosome. (4) GEls are often flanked by direct repeats that may have been generated during integration of GEl-specific regions into the host chromosome via site-specific recombination. (5) GEls are frequently associated with tRNA loci. The 3' end oftRNA genes have heen recognized as preferred target sites for the integration offoreign DNA [reviewed in I I]. (6) GEls often possess functional or cryptic genes coding for factors that are involved in genetic mobility such as integrases, transposases, phage genes and origins of replication. Furthermore, GEls normally do not represent homogenous elements but

Pathogenicity Islands

235

rather are generated by multistep processes including DNA rearrangements via IS elements which is reflected by mosaic-like structures. (7) Some GEls tend to be unstable DNA regions due to recombination between the flanking direct repeats, between IS elements or between other regions of homologous sequences. Generally, little is known about the mechanisms that have led to the acquisition of GEls and there are only few examples of inter- or intracellular mobilization of GEls [12-16]. GEls are prevalently found in organisms that show frequent gene transfer by bacteriophages and plasmids which are regarded as possible precursors of GEls [8]. However, GEls have also been described in bacteria that exhibit natural competence such as Helicobacter pylori, Neisseria gonorrhoeae and Streptococcus pneumoniae, and that tend to introduce smaller pieces rather than large regions of foreign DNA into their genome [17-19].

GEls Contribute to Bacterial Fitness Besides selfish genes such as genes involved in recombination and transfer or modification of DNA, GEls often carry determinants that are beneficial for their host bacterium in certain environments thereby increasing bacterial fitness and consequently survival. GEls were divided into different subtypes reflecting their contribution to the respective microbial lifestyle [8] (table 1). GEls that encode virulence traits were defined as 'pathogenicity islands' (PAIs). The original definition of GEls was based on the characteristics of PAls in pathogenic E. coli, but intensive studies of the genome structure of bacterial pathogens resulted in the identification of similar structures in many phylogenetically unrelated organisms including gram-negative as well as gram-positive bacteria (tables 2-3). Typical virulence factors encoded on PAIs include toxins, adhesins and fimbriae, factors involved in host cell entry, capsules, secretion systems and iron uptake systems. Based on the broad distribution of PAIs, it can be concluded that they have contributed significantly to the evolution of virulent variants. However, the still growing number of genome sequences has made it clear that GEls are not restricted to pathogenic species. GEls contributing to the adaptalion lo specific growlh condilions or the inleraclion wilh a eukaryolic host organism have been described in environmental, commensal or symbiotic bacteria and have been designated 'symbiosis islands', 'ecological islands' or 'resistance islands', according to the respective encoded functions. Relatively well-studied examples of GEls include the symbiosis island of Mesorhizobium melioti that carries genes required for nitrogen fixation, whereas GEls such as the mec region enhance survival of staphylococci in hospitals where they have to face antimicrobial substances. Other islands encode enzymes involved in the degradation of phenolic compounds or for uptake and metabolism of certain carbohydrates (table I). Finally, a recently described island in Magnetospirillum

HochhutiDobrindtiHacker

236

Table 1. Examples of GEls

Subtype of

Designation

Organism

Encoded functions

Reference

PAl

PAIll5J6

Escherichia coli

P fimbriae,